IESNA Lighting Handbook - PDF Free Download (2023)

INDEX

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

-

Light and Optics Measurement of Light and Other Radiant Energy Vision and Perception Color Nonvisual Effects of Optical Radiation Light Sources Luminaires Daylighting Lighting Calculations Quality of the Visual Environment Office Lighting Educational Facility Lighting Hospitality Facilities and Entertainment Lighting Lighting for Public Places and Institutions Theatre,Television and Photographic Lighting Health Care Facility Lighting Retail Lighting Residential Lighting Industrial Lighting Sports and Recreational Area Lighting Exterior Lighting Roadway Lighting Transportation Lighting Underwater Lighting Lighting Economics Energy Management Lighting Controls Lighting Maintenance Emergency, Safety, and Security Lighting Appendix Glossary of Lighting Terminology

IESNA Illuminating Engineering Society of North America The IESNA is the recognized technical authority on illumination. For over ninety years its objective has been to communicate information on all aspects of good lighting practice to its members, to the lighting community, and to consumers through a variety of programs, publications, and services. The strength of the IESNA is its diversified membership: engineers, architects, designers, educators, students, contractors, distributors, utility personnel, manufacturers, and scientists, all contributing to the mission of the Society: to advance knowledge and disseminate information for the improvement of the lighted environment to the benefit of society. The IESNA is a forum for the exchange of ideas and information and a vehicle for its members' professional development and recognition. Through its technical committees, with hundreds of qualified members from the lighting and user communities, the IESNA correlates research, investigations, and discussions to guide lighting experts and laypersons via consensus-based lighting recommendations. The Society publishes nearly 100 varied publications including recommended practices on a variety of applications, design guides, technical memoranda, and publications on energy management and lighting measurement. The Society, in addition, works cooperatively with related organizations on a variety of programs and in the production of jointly published documents and standards. In addition, the Society publishes Lighting Design + Application (LD+A) and the Journal of the Illuminating Engineering Society (JIES). LD+A is a popular application-oriented monthly magazine. Every issue contains special feature articles and news of practical and innovative lighting layouts, systems, equipment and economics, and news of the industry. The Journal contains technical papers, most of which are presented at the Society's Annual Conference. IESNA has a strong education program with basic and intermediate level courses and seminars offered through its Sections. The Society has two types of membership: individual and sustaining. Applications and current dues schedules are available upon request from the Membership Department. IESNA local, regional, and international meetings, conferences, symposia, seminars, workshops, and lighting exhibitions (LIGHTFAIR INTERNATIONAL) provide current information on the latest developments in illumination. For additional information on the IESNA, consult the Society's Web site: <www.iesna.org>.

TUMINATING ENGINEERING SOCIETY OF NORTH AMERICA Managing Editor: Judith Block Production Manager: Judith Block Editorial Assistants: John Bullough, Mariana Figueiro, and Marilyn R. P. Morgan Copyeditor: Seth A. Maislin Illustrator: Joseph R. Gilmore Indexer: Specialized Scientific Indexing Typesetting: Eastern Composition Marketing: PamelaWeess Cover Design: Tony Picco The IESNA LIGHTING HANDBOOK, Ninth Edition

Copyright © 2000 by the Illuminating Engineering Society of North America. All rights reserved. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by IESNA provided that the base fee of $5.00 per copy plus $2.00 per page per copy is paid directly to CCC, 27 Congress Street, Salem, MA 01970. When submitting payment please indicate the publication material was taken from, the page numbers, and the ISBN of the publication. This consent does not extend to any other kinds of copying and the publication may not be duplicated in any other way without the express written consent of the publisher. This includes, but is not limited to, duplication in other publications, databases or any other medium for purposes of general distribution for resale. Making copies of this book, or any portion for any purposes other than personal use, is a violation of United States copyright laws and will be subject to penalty. ISBN 0-87995-150-8 Library of Congress Catalog Card Number: 99-76610. Printed in the United States of America. The Illuminating Engineering Society of North America welcomes your comments. Please send all correspondence to: Publications Department IESNA 120 Wall Street, 17th Floor New York, NY 10005-4001

Preface Many of us believe that the ninth edition of the IESNA Lighting Handbook represents a watershed in lighting practice. Over the past twenty years there has been a movement in lighting practice from illuminating engineering to lighting design, a movement from calculations of illuminance to judgments of aesthetics, a movement from quantity to quality. For the first time, the IESNA has, through this edition, formalized recommendations of lighting quality, reflecting this movement in lighting practice. These formal recommendations are provided in a matrix entitled the IESNA Lighting Design Guide. The Guide includes recommendations on important lighting design criteria such as eye-source-task geometry, flicker, color, and glare. They are provided alongside the traditional recommendations of illuminance for a wide variety of applications. The intent of the Guide is to broaden the perspective of lighting practitioners and to direct them to specify higher quality lighting. The idea for the IESNA Lighting Design Guide was born beside Lake George in upstate New York at the retreat for editing the eighth edition of this Handbook. During a break in editing, some of the editing team took a walk along the edge of the lake. Feeling a bit tired, we lamented that most people would probably never read what we were editing because they would only consult the illuminance selection table. We repeated the standard joke: If the Handbook in most architectural-engineering offices is placed on its spine, it will fall open to the illuminance selection table because that is the only section ever consulted. We mused that it be nice if users had to consider the many other important lighting design criteria found throughout the text. Building on that idea, we sketched out the basic framework of the matrix. Through hard work and review by several committees, the IESNA Lighting Design Guide in Chapter 10 of this

edition of the Handbook was produced. Of course we all still hope that users will read the entire text of the ninth edition of the Handbook, but if it must fall open to any one section, it will now fall open to a section that describes more than one lighting design criterion. Actually the genesis of the IESNA Lighting Design Guide goes back several years before our walk along Lake George. One of the first lighting people I met as a graduate student at Ohio State University was Steve Squillace, engineer, teacher, past president of the IESNA, and recipient of its highest technical award. Steve's passion for lighting and life set him apart from his contemporaries. In every conversation I had with Steve he insisted that every lighting designer and illuminating engineer should think about lighting. As a passionate radical, he argued that the IESNA should do away with illuminance recommendations altogether because they were substitutes for thinking. Many disagreed with Steve, believing that the majority of practitioners in the building industry were not lighting specialists. These people needed to quickly find practical guidance and then move on to other decisions. The problem had been that illuminance became the only criterion these practitioners considered before moving on. In this edition, the single focus on illuminance is no longer possible. Steve might argue that we are moving in the wrong direction, however. There are more formal recommendations in this edition of the Handbook than ever before. Perhaps people who are now doing good lighting will stop thinking, but we doubt that. Rather, we believe that thinking is required to follow these recommendations. The lighting practitioner who needs to hurry to the next decision can no longer rely upon an illuminance calculation and consider the lighting job completed. Often illuminance is not the primary lighting design criterion in the Guide. With the recommendations put forward in this edition, the practitioner must take some time to study the application and decide among several important lighting design criteria. The thinking time invested by the lighting practitioner is worthwhile because that investment will improve the quality of lighting throughout North America. Many people deserve a great deal of credit in developing, writing, and producing this edition of the IESNA Lighting Handbook. I have tried to acknowledge everyone who contributed, but no acknowledgment can do justice to the long-standing commitment these people have made to lighting. Their contributions to this edition are only a small part of their life-long commitment to improving the quality of life through better lighting. It is my sincerest wish that the ninth edition of the IESNA Lighting Handbook does honor to these contributors and helps them continue to improve the quality of lighting throughout North America. Mark S. Rea, Ph.D., FIES Editor-in-chief

Foreword The Illuminating Engineering Society was founded in 1906, but it was not until 1947 that the first edition of the Handbook appeared, thus representing the accumulation of 41 years of lighting progress since the Society's founding. In each subsequent edition, IESNA has provided information on an ever-broadening range of technologies, procedures, and design issues. In the ninth edition, the editorial team has continued the trend of securing knowledge on all phases of lighting from IESNA committees and individual experts to ensure that this Handbook is the lighting reference source for the beginning of the next century. The emphasis in the ninth edition is on quality. Previous editions have discussed important criteria for assessing and designing the visual environment, but a formal system for considering these issues had not been developed. IESNA has, however, always recommended quantity of light for specific applications or visual tasks. As a result, many practitioners often mistook the IESNA system of recommended illuminances (quantity) as the primary, or even the sole criterion, for lighting design. This Handbook introduces a new, formal system of addressing quality issues in the Lighting Design Guide in Chapter 10,

Quality of the Visual Environment. There are changes, too, in the illluminance categories, reduced from nine to seven and organized into three sets of visual tasks (simple, common, and special). Every application in the Lighting Design Guide has a specific (single number) recommended illuminance representing best practice for a typical application. Through the Lighting Design Guide and other information in Chapter 10, IESNA is recognizing and emphasizing that illuminance is not the sole lighting design criterion. Other criteria may be more important, and, given the complexity and diversity of design goals for a specific application, the designer now has the opportunity to evaluate among the quantity and quality choices. This approach has been described as "a bridge to the 21st century," when it is expected that the tenth edition of the Handbook will provide a more precise method of measuring quality factors and their impact on the visual environment. Other chapters in the book are new, or have been rewritten or updated. There are new application chapters on outdoor lighting, security lighting, parking facilities, retail, shopping mall and industrial lighting, and significant revisions to chapters on measurement of light, vision and perception, photobiology, aviation, and transportation. This Handbook could not have been produced without the IESNA committees and individual specialists, those willing volunteers who give countless hours to the process of sharing their expertise. The Society thanks each and every contributor. The professional editorial team brought talent and discipline to the project. Dr. Mark Rea, Judith Block, John Bullough, and Mariana Figueiro of Rensselaer Polytechnic Institute together with four Topic Editors, Michael Ouellette, David DiLaura, Roger Knott, and Nancy Clanton, have earned our appreciation for their contributions in evaluating, editing and, when necessary, developing material. The IESNA Lighting Handbook represents the most important reference document in the lighting profession. It is one means by which the Society accomplishes its mission: to advance knowledge and disseminate information for the improvement of the lighted environment to the benefit of society. We hope that, you, the reader, will find the ninth edition your principal reference source for lighting information. William H. Hanley Executive Vice President Rita M. Harrold Director, Educational and Technical Development

Acknowledgments We acknowledge the four Topic Editors: Nancy E. Clanton, Clanton Engineering Associates, Boulder, Colorado David DiLaura, University of Colorado at Boulder Roger Knott, Lighting Consultant, Cleveland, Ohio Michael Ouellette, National Research Council Canada, Ottawa

We acknowledge the following committees and committee chairs for their efforts on behalf of this revision of the Handbook during the period 19961999: Agricultural Lighting: Ronald MacDonald, Chair Aviation: William Pickell, Chair (199697), Daniel Geary, Chair (199799) Calculation Procedures: Richard G. Mistrick, Chair (199697), Ian Ashdown, Chair (199799) Casino and Gaming Lighting: Elwyn Gee, Chair Color: Ron Daubach, Chair Computer: Paul K. Ericson, Chair Correctional Facilities: Stewart E. Greene, Chair Daylighting: Morad R. Atif, Ph.D.,Chair Emergency Lighting: Mary Kim Reitterer, Chair Energy Management: Dave Ranieri, Chair (199698), Carol Jones, Chair (199899) Financial Facilities: Hyman M. Kaplan, Chair (199698) Health Care Facilities: David H. Epley, Chair Hospitality Facilities: Candace M. Kling, Chair (199697), Martyn Timmings, Chair (199799) Houses of Worship: Viggo B. Rambusch, Chair Industrial: William T. Busch, Chair Landscape Lighting, Lloyd Reeder, Chair Light Control and Luminaire Design: Gerald Plank Jr., Chair Light Sources: Pekka Hakkarainen, Chair Lighting Economics: John Selander, Chair (199697), Cheryl English, Chair (199799) Lighting for the Aged and Partially Sighted: Eunice D. Noell, Chair Maintenance: Norma Frank, Chair Mall Lighting: Robert Horner, Chair Marine Lighting: Michael J. Leite, Chair Museum and Art Gallery: Frank A. Florentine, Chair Nomenclature: Warren "Gus" Baker, Chair Office Lighting: Mitchell B. Kohn, Chair Outdoor Environmental: Nancy E. Clanton, Chair Photobiology: George C. Brainard, Ph.D., Chair Quality of the Visual Environment: Naomi Johnson Miller, Chair Residence Lighting: Kathy A. Presciano, Chair Retail Areas: Bernie Bauer, Chair Roadway: Balu Ananthanarayanan, Chair (199698), John Mickel, Chair (199899) School and College: Shail Mahanti, Chair Security Lighting: Douglas Paulin, Chair (199698), David Salmon, Chair (199899) Sports and Recreational Areas: John Kirchner, Chair (199698), Michael Owens (199899) Technical Review Council: Donald Smith, Chair (199697), Richard Collins, Chair (199799) Testing Procedures: Richard Collins, Chair (199697), James Walker, Chair (199799) Theatre, Television, and Film: James P. McHugh, Chair Contributing individuals, in alphabetical order: Eric Block, Peter Boyce, Jack Burkarth, Christopher Cuttle, David Evans, Joseph M. Good, III, Dawn DeGrazio, Jim Fowler, Rita M. Harrold, Hugh Henry, Jules Horton, Jules S. Jaffe, Yunfen Ji, Walter J. Kosmatka, Robert Landry, Robert Levin, Kevin McCarthy, Greg McKee, Nishantha Maliyagoda, Scott Mangum, Naomi Johnson Miller, Sharon Miller, Janet Lennox Moyer, Joseph Murdoch, N. Narendran, Peter Ngai, Yoshihiro Ohno, Mark Olsson, Robert Roller, Greg Shick, Ted Smith, Stephen Squillace, Gary Steffy, Jennifer Veitch.

1 Light and Optics The quest to understand the nature of light has led curious human beings down into the innermost secrets of the atom and out to the farthest reaches of the starry universe. --Ben Bova

FUNDAMENTALS For illuminating engineering purposes, the Illuminating Engineering Society of North America (IESNA) defines light as radiant energy that is capable of exciting the human retina and creating a visual sensation. As a physical quantity, light is defined in terms of its relative efficiency throughout the electromagnetic spectrum lying between approximately 380 and 780 nm. Visually, there is some individual variation in efficiency within these limits.

Theories One of the earliest theories to describe light involved the notion that light was emitted from the eyes, and that they were rendered visible when they were struck by the emissions. Aristotle rejected this theory when questioning why we could not see in the dark. Since then, many alternative theories have been advanced. From a physical point of view, these theories generally regarded light as an energy transfer from one location to another. Some theories1-4 are briefly discussed below. Corpuscular Theory. This theory follows from the observation that moving particles, or corpuscles, possess kinetic energy. This position was advocated by Sir Isaac Newton (1642-1727). It is based on three premises: 1. Luminous bodies emit radiant energy in particles. 2. The particles are intermittently ejected in straight lines. 3. The particles act on the retina, stimulating a response that produces a visual sensation. Wave Theory. This theory follows from the observation that waves can transfer energy even though the medium itself does not travel. This position was advocated by Christiaan Huygens (1629-1695). It too is based on three premises: 1. Light results from the molecular vibration in the luminous material. 2. The vibrations are transmitted through an "ether" as wavelike movements (comparable to ripples in water), and the vibrations slow down upon entering denser media. 3. The transmitted vibrations act on the retina, stimulating a response that produces a visual sensation. Electromagnetic Theory.5 The theory was advanced by James Clerk Maxwell (1831-1879), and is based on three premises: 1. Luminous bodies emit light in the form of radiant energy. 2. Radiant energy is propagated in the form of electromagnetic waves. 3. The electromagnetic waves act upon the retina, stimulating a response that produces a visual sensation. Quantum Theory. A modern form of the corpuscular theory was advanced by Max Planck, and is based on two premises: 1. Energy is emitted and absorbed in discrete quanta (photons).

2. The magnitude of each quantum, Q, is determined by the product of h and ν, where h is 6.626 × 10−34 J·s (Planck's constant), ν is the frequency of the photon vibration in Hz, and Q is energy in Joules. This theory provides a means of determining the amount of energy in each quantum. It follows from this theory that energy increases with frequency. Unified Theory. The theory proposed by Louis de Broglie and Werner Heisenberg is based on two premises: 1. Every moving element of mass has associated with it a wave whose length is given by the equation

where λ = wavelength of the wave motion, h = Planck's constant, m = mass of the particle, v = velocity of the particle.

Figure 1-1. The radiant energy (electromagnetic) spectrum. 2. It is impossible to simultaneously determine all of the properties that are distinctive of a wave or a corpuscle. The quantum and electromagnetic wave theories provide an explanation of those characteristics of radiant energy of concern to the illuminating engineer. Whether light is considered as a wave or a photon, it is radiation that is produced by electronic processes in the most exact sense of the term. It is produced in an incandescent body, a gas discharge, or a solid-state device by excited electrons just having reverted to more stable positions in their atoms, releasing energy.

Light and the Energy Spectrum 6 The wave theory permits a convenient graphical representation of radiant energy in an orderly arrangement according to its wavelength or frequency. This arrangement is called a spectrum (Figure 1-1). It is useful in indicating the relationship between various radiant energy wavelength regions. Such a graphical representation should not be construed to indicate that each region of the spectrum is divided from the others in any physical way; there is a gradual transition from one region to another. The radiant energy spectrum extends over a range of wavelengths from 10−16 to 105 m. The Angstrom unit (Å), the nanometer (nm), and the micrometer (µm), which are respectively 10−10, 10−9, and 10−6 m, are commonly used units of length in the visible spectrum region. The nanometer is the preferred unit of wavelength in the ultraviolet (UV) and visible regions of the spectrum. The micrometer is normally used in the infrared (IR) region.

Of particular importance to illuminating engineering are three regions of the electromagnetic spectrum: UV, visible, and IR. On the basis of practical applications and the effect obtained, the UV region is divided into the following bands (for engineering purposes, the "black light" region extends slightly into the visible portion of the spectrum):

Another division of the UV spectrum, often used by photobiologists, is given by the Commission Internationale de l'Éclairage (CIE):

Radiant energy in the visible spectrum lies between 380 and 780 nm. For practical purposes, infrared radiant energy is within the wavelength range of 0.78 to 103 µm. This band is arbitrarily divided as follows:

In general, unlike UV energy, IR energy is not evaluated on a wavelength basis but rather in terms of all such energy incident upon a surface. Examples of these applications are industrial heating, drying, baking, and photoreproduction. However, some applications, such as IR viewing devices, involve detectors sensitive to a restricted range of wavelengths; in such cases the spectral characteristics of the source and receiver are of importance.

Figure 1-2. Speed of Light for a Wavelength of 589 nm (Na D-lines) All forms of radiant energy are transmitted at the same speed in vacuum (299,793 km/s, or 186,282 mi/s). However, each form differs in wavelength and thus in frequency. The wavelength and velocity may be altered by the medium through which it passes, but the frequency remains constant, independent of the medium. Thus, through the equation

where n = index of refraction of the medium, λ = wavelength in a vacuum, ν = frequency in Hz, it is possible to determine the velocity of radiant energy and also to indicate the relationship between frequency and wavelength. Figure 1-2 gives the speed of light in different media for a frequency corresponding to a wavelength of 589 nm in air. Light is . . . a certain motion or an action, conceived in a very subtle manner, which fills the pores of all other bodies. . . . --René Descartes, in La Dioptrique, 1637

Blackbody Radiation

The intensity and spectral properties of a blackbody radiator are dependent solely upon its temperature. A blackbody radiator may be closely approximated by the radiant power emitted from a small aperture in an enclosure, the walls of which are maintained at a uniform temperature (Figure 1-3). Emitted light from a practical light source, particularly from an incandescent lamp, is often described by comparison with that from a blackbody radiator. In theory, all of the energy emitted by the walls of the blackbody radiator is eventually reabsorbed by the walls; that is, none escapes from the enclosure. Thus, a blackbody will, for the same area, radiate more total power and more power at a given wavelength than any other light source operating at the same temperature.

Figure 1-3. Small aperture in an enclosure exhibits blackbody characteristics. From 1948 to 1979, the luminance of a blackbody, operated at the temperature of freezing platinum (2042 K), was used as an international reference standard to define the unit of luminous intensity. Specifically, it has a luminance of 60 cd/m2. Since operating and maintaining a blackbody radiator at the freezing point of platinum is a major undertaking, a new definition of the candela was adopted in 1979. The candela is now, essentially, the luminous intensity of a 555.016 nm source whose radiant intensity is 1/683 W/sr. The new photometric unit is based on an electrical unit, the watt, which can be accurately and conveniently measured with an electrically calibrated radiometer. A further advantage of this definition is that the magnitude of the unit is independent of the international temperature scale, which occasionally changes. Planck Radiation Law. Data describing blackbody radiation curves were obtained by Lummer and Pringsheim using a specially constructed and uniformly heated tube as the source. Planck, introducing the concept of discrete quanta of energy, developed an equation depicting these curves. It gives the spectral radiance of a blackbody as a function of wavelength and temperature. See the definition of Planck's radiation law in the Glossary. Figure 1-4 shows the spectral radiance of a blackbody, on a logarithmic scale, as a function of wavelength for several absolute temperatures. Wien Radiation Law. In the temperature range of incandescent filament lamps (2000 to 3400 K) and in the visible wavelength region (380 to 780 nm), a simplification of the Planck equation, known as the Wien radiation law, gives a good representation of the blackbody distribution of spectral radiance (see the Glossary). Wien Displacement Law. This gives the relationship between the peak wavelength of blackbody radiation at different temperatures (see line AB in Figure 1-4, and the Glossary). Stefan-Boltzmann Law. This law, obtained by integrating Planck's expression for Lλ from zero to infinity, states that the total radiant power per unit area of a blackbody varies as the fourth power of the absolute temperature (see the Glossary). It should be noted that this law applies to the total power, that is, the whole spectrum. It cannot be used to estimate the power in the visible portion of the spectrum alone.

Figure 1-4. Blackbody radiation curves for operating temperatures between 500 and 20,000 K, showing Wein displacement of peaks. The shaded area is the region of visible wavelengths.

Spectral Emissivity No known radiator has the same emissive power as a blackbody. The ratio of the output of a radiator at any wavelength to that of a blackbody at the same temperature and the same wavelength is known as the spectral emissivity, ε(λ), of the radiator.

Graybody Radiation When the spectral emissivity is constant for all wavelengths, the radiator is known as a graybody. No known radiator has a constant spectral emissivity for all visible, IR, and UV wavelengths, but in the visible region a carbon filament exhibits nearly uniform emissivity; that is, a carbon filament is nearly a graybody for this region of the electromagnetic spectrum.

Selective Radiators When the emissivity of all known material varies with wavelength, the radiator is called a selective radiator. In Figure 15, the radiation curves for a blackbody, a graybody, and a selective radiator (tungsten), all operating at 3000 K, are plotted on the same logarithmic scale to show the characteristic differences in radiant power.

Figure 1-5. Radiation curves for blackbody, graybody, and selective radiators operating at 3000 K.

Color Temperature and Distribution Temperature The radiation characteristics of a blackbody of unknown area may be specified with the aid of the Planck's equation by fixing only two quantities: the magnitude of the radiation at any given wavelength, and the absolute temperature. The same type of specification may be used with reasonable accuracy in the visible region of the spectrum for tungsten filaments and other incandescent sources. However, the temperature used in the case of selective radiators is not that of the filament but a value called the color temperature. The color temperature of a selective radiator is that temperature at which a blackbody would have to be operated to produce the same color as that of the selective radiator. Color temperature is calculated from the chromaticity coordinates (u,v) of the source; small differences between chromaticities of a blackbody and an incandescent filament lamp are not of practical importance. This is true because the interreflections that occur at the inner surfaces of the helix formed by the coils used in many tungsten lamps act somewhat like a blackbody radiator. Thus, the spectral power distributions from coiled filaments exhibit combined characteristics of a straight filament and of a blackbody operating at the same temperature. Distribution temperature is the temperature of a blackbody whose relative spectral power distribution is the closest to that of the given selective radiator. Distribution temperature is defined from the spectral power distribution of the source.7

Figure 1-6. Schematic structure of the atom, showing electron orbits around a central nucleus. Hydrogen and helium atoms are the simplest of all atomic structures. Color temperature and distribution temperature apply only to incandescent sources. Correlated color temperatures are

used to describe the light emitted from other types of sources. Color is discussed in greater detail in Chapter 4, Color.

Atomic Structure and Radiation The atomic theories first proposed by Rutherford and Bohr in 1913 have since been expanded upon and confirmed by an overwhelming amount of experimental evidence. They hypothesize that each atom resembles a minute solar system, such as that shown in Figure 1-6. The atom consists of a central nucleus possessing a positive charge +n, about which revolve n negatively charged electrons. In the normal state these electrons remain in particular orbits, or energy levels, and radiation is not emitted by the atom. The orbit described by a particular electron rotating about the nucleus is determined by the energy of that electron. In other words, there is a particular energy associated with each orbit. The system of orbits or energy levels is characteristic of each element and remains stable unless disturbed by external forces. The electrons of an atom can be divided into two classes. The first class includes the inner shell electrons, which are not readily removed or excited except by high-energy radiation. The second class includes the outer shell (valence) electrons, which cause chemical bonding into molecules. Valence electrons are readily excited by UV or visible radiation or by electron impact and can be removed with relative ease. The valence electrons of an atom in a solid, when removed from their associated nuclei, enter the so-called conduction band and confer on the solid the property of electrical conductivity. Upon the absorption of sufficient energy by an atom in the gaseous state, the valence electron is pushed to a higher energy level further from the nucleus. Eventually, the electron returns to the normal orbit, or an intermediate one, and in so doing the energy that the atom loses is emitted as a quantum of radiation. The wavelength of the radiation is determined by Planck's formula:

where E2 = energy associated with the excited orbit, E1 = energy associated with the normal orbit, h = Planck's constant, ν21 = frequency of the emitted radiation as the electron moves from level 2 to level 1. This formula can be converted to a more usable form:

where Vd = potential difference in volts between two energy levels through which the displaced electron has fallen in one transition.

Luminous Flux and the Lumen8,9 Of particular importance to illuminating engineering is the lumen. The goal of this section is to show how electric power of radiant flux (in watts) is converted into luminous flux (in lumens), and to describe the underlying rationale for this process. The lumen is, in fact, a unit relating radiant flux (in watts) to visually effective radiation (i.e., light) for a standard human observer. There are two classes of photoreceptors in the human eye, rods and cones. The photopic function Vλ describes the spectral luminous efficiency function for photopic (cone) vision, and the scotopic function V'λ describes the spectral luminous efficiency for scotopic (rod) vision (Figure 1-7).10

The photopic luminous efficiency function Vλ was established in 1924 by Commission Internationale de l'Éclairage (CIE) and is based on data from several experimenters using different techniques. The two primary techniques used were flicker photometry and step-by-step heterochromatic brightness matching. Flicker photometry is the least variable technique for determining the photopic efficiency function. With this technique, two lights are seen alternately in rapid succession. The radiance of one light, called the reference light, is held constant while the radiance of the other light, called the test light, which is monochromatic, is varied to the point where minimum flicker is perceived. At this point the luminances of both lights are defined to be the equal. Each test light wavelength is compared with the reference light in this way. The wavelength associated with the reciprocal of the minimum radiance needed to match the reference light is defined as the unit value of the photopic spectral luminous function (Vλ =1). In heterochromatic brightness matching, the reference light of constant radiance is juxtaposed with the test wavelength of variable radiance. The subject simply adjusts the radiance of each test wavelength until it appears to be equal in brightness to the reference. This technique is highly variable and produces results very different from flicker photometry unless the spectral difference between the test wavelength and the reference light is small. To obtain useful results, then, the reference light must be different for different regions of the spectrum. Since the reference light changes across the spectrum, this method is known as the step-by-step heterochromatic brightness matching technique. Again, the wavelength associated with the minimum value needed to match the reference light(s) is defined as the unit value of the photopic function. Several consistent experimental conditions were used in these early experiments. The test fields were small, usually less than 2° across; the luminance was fairly low due to light source limitations, and a natural pupil was used by the subjects during testing. Gibson and Tyndall11 pieced together results from several experiments and recommended a particular spectral luminous efficiency function for the photopic (cone) system, which was approved by a committee of the CIE in 1924. Modification to the CIE 1924 curve followed, based on work by Judd in 1951. The 1924 curve was shown to be inadequate in describing visual sensitivity in the short-wavelength region of the visible spectrum. This modified curve was later published by the CIE.12 Since not all test fields of interest to experimentalists were 2° or less, a standard function for a 10° field was devised in 1964, which shows a still greater sensitivity to short wavelengths on the photopic curve. This is likely due to the macula lutea screening pigment (see Chapter 3, Vision and Perception). In 1951 the CIE also established a scotopic luminous efficiency function (Figure 1-7) based on the heterochromatic brightness matching technique (not step-by-step). Test wavelengths were compared with a large, approximately 20°, "white" test field with a luminance of approximately 0.00003 cd/m2. The field was viewed by subjects using natural pupils after an extended period in darkness (see also Figure 3-8). It is important to point out that everyone in the previous studies was color normal. A small percentage of the population (approximately 8%, mostly males) do not have all three cone photopigments or do not have the same ones as colornormal people. The photopic luminous efficiency curves will be different for these people because the cone photopigments determine the shapes of these curves. The photopic luminous efficiency function applies to visual stimuli to the fovea and at luminance levels higher than approximately 3 cd/m2. The scotopic luminous efficiency function applies to visual stimuli in regions outside the fovea and to luminances below approximately 0.001 cd/m2. A family of mesopic luminous efficiency functions is required for application to luminous stimuli between approximately 0.001 and 3 cd/m2. Research in this area is on-going.13,14 Presently, mesopic luminous efficiency functions remain to be defined officially.15 With the exception of special measurements for research purposes, almost all photometric quantities are measured photopically, even at luminances below 3 cd/m2 and for peripheral vision. See Chapter 3, Vision and Perception, for additional discussion on photopic, scotopic, and mesopic vision.

Luminous Efficacy of Light Sources The luminous efficacy of a light source is defined as the ratio of the total luminous flux (in lumens) to the total power input (in watts). There are 683 lumens/watt at 555 nm. Since the scotopic luminous efficiency function peaks at a different wavelength (507 nm), it is necessary to establish different scaling factors for the photopic and for the scotopic luminous efficiency functions. Therefore, the photopic lumens, F, and the scotopic lumens, F ', must be determined from the spectral power distribution of the light source:

Figure 1-7. Photopic Luminous Efficiency, V (λ), and Scotopic Luminous Efficiency, V'(λ) Functions

where Pλ = spectral power, in watts, of the source at the wavelength λ, Vλ = photopic luminous efficiency function value at λ, ∆λ = interval over which values of the spectral power were measured, and

where Vλ' = scotopic luminous efficiency function value at λ. The maximum luminous efficacy of an ideal white source, defined as a radiator with constant output over the visible part of the spectrum and no radiation in other parts, is approximately 220 lm/W.

LIGHT GENERATION Natural Phenomena Sunlight. Energy with a color temperature of approximately 6500 K is received from the sun just outside the earth's atmosphere at an average rate of about 1350 W/m2. About 75% of this energy reaches the earth's surface at sea level (on the equator) on a clear day. The average luminance of the sun is approximately 1600 Mcd/m2 viewed from sea level. The illuminance on the earth's surface by the sun may exceed 100 klx (10,000 fc); on cloudy days the illuminance drops to less than 10 klx (1000 fc). Formulas to calculate these values are in Chapter 8, Daylighting. Sky Light. A considerable amount of light is scattered by the earth's atmosphere. The investigations of Rayleigh first showed that this was a true scattering effect. On theoretical grounds the scattering should vary inversely as the fourth power of the wavelength when the size of the scattering particles is small compared to the wavelength of light, as in the case of the air molecules themselves. The blue color of a clear sky and the reddish appearance of the rising or setting sun are common examples of this scattering effect. If the scattering particles are relatively large (the water droplets in a cloud, for example), scattering is essentially the same for all wavelengths (clouds appear white). The scattered light from parts of the sky is partially polarized, up to 50%. Moonlight. The moon shines solely by reflection of sunlight. Since the reflectance of its surface is rather low, its luminance is only on the order of 2500 cd/m2. The correlated color temperature of moonlight is around 4100 K but will vary widely depending on material suspended in the atmosphere. Illumination of the earth's surface by the moon can be as high as 0.1 lx (0.01 fc). Lightning. Lightning is a meteorological phenomenon arising from the accumulation, in the formation of clouds, of tremendous electrical charges, usually positive, which are suddenly released in a spark discharge. The lightning spectrum corresponds closely to that of an ordinary spark in air, consisting principally of nitrogen bands, although hydrogen lines sometimes appear owing to dissociation of water vapor. Aurora Borealis (Northern Lights) and Aurora Australis (Southern Lights). These hazy patches or bands of greenish light, on which white, pink, or red streamers sometimes are superposed, appear 100 to 200 km (60 to 120 mi) above the earth. They are caused by electron streams spiraling into the atmosphere, primarily at polar latitudes. Some of the lines in their spectra have been identified with transitions of valence electrons from metastable states of oxygen and nitrogen atoms. Bioluminescence. "Living light" is a form of chemiluminescence in which special compounds manufactured by plants and animals are oxidized, producing light. The light-producing compounds are not always required to be in a living organism. Many bioluminescent compounds can be dried and stored many years and then, in response to exposure to oxygen or some other catalyst, emit light.

Fabricated Sources Historically, light sources have been divided into two types, incandescent and luminescent. Fundamentally, the cause of light emission is the same: electronic transitions from higher to lower energy states. The mode of electron excitation and the resultant spectral distribution of the radiation are different, however. Incandescent solid substances emit a continuous spectrum, while gaseous discharges radiate mainly in discrete spectral lines. There is some overlap, however. Incandescent rare-earth elements can emit discrete spectra, whereas high-pressure discharges produce a continuous spectrum. The two classic types, with subdivisions showing associated devices or processes, are listed as follows (see also Chapter 6, Light Sources, for discussion on some of the following): I. Incandescence

A. Filament lamps B. Pyroluminescence (flames) C. Candoluminescence (gas mantle) D. Carbon arc radiation II. Luminescence A. Photoluminescence 1. Gaseous discharges 2. Fluorescence 3. Phosphorescence 4. Lasers B. Electroluminescence 1. Electroluminescent lamps (ac capacitive) 2. Light-emitting diodes 3. Cathodoluminescence (electron excitation) C. Miscellaneous luminescence phenomena 1. Galvanoluminescence (chemical) 2. Crystalloluminescence (crystallization) 3. Chemiluminescence (oxidation) 4. Thermoluminescence (heat) 5. Triboluminescence (friction or fracture) 6. Sonoluminescence (ultrasonics) 7. Radioluminescence (α, β, γ, and X rays)

INCANDESCENCE Incandescent Filament Lamps All familiar physical objects are combinations of chemically identifiable molecules, which in turn are made up of atoms. In solid materials the molecules are packed together, and the substances hold their shape over a wide range of physical conditions. In contrast, the molecules of a gas are highly mobile and occupy only a small part of the space filled by the gas. Molecules of both gases and solids are constantly in motion at temperatures above absolute zero (0 K or 273°C), and their movement is a function of temperature. If the solid or gas is hot, the molecules move rapidly; if it is cold, they move more slowly. At temperatures below approximately 873 K (600°C), only IR energy (heat) is emitted by a body, for example, a coal stove or an electric iron. Electronic transitions in atoms and molecules at temperatures above approximately 600°C result in the release of visible radiation along with the heat. The incandescence of a lamp filament is caused by the heating action of an electric current. This heating action raises the filament temperature substantially above 600°C, producing light.

Pyroluminescence (Flame Luminescence) A flame is the most often noted visible evidence of combustion. Flame light may be due to recombination of ions to form molecules, reflection from solid particles in the flame, incandescence of carbon or other solid particles, or any combination of these. The combustion process is a high-temperature energy exchange between highly excited molecules and atoms. The process releases and radiates energy, some of which is in that portion of the electromagnetic spectrum called light. The quality and the amount of light generated depend on the material undergoing combustion. For example, a flashbulb containing zirconium yields the equivalent of 56 lm/W, whereas an acetylene flame yields 0.2 lm/W.

Candoluminescence (Gas Mantle)

Incandescence is exhibited by heated bodies which give off shorter wavelength radiation than would be expected according to the radiation laws, because of fluorescence excited by incandescent radiation. Materials producing such emission include zinc oxide, as well as rare-earth elements (cerium, thorium) used in the Welsbach gas mantle.

Carbon Arc Radiation A carbon arc source radiates because of incandescence of the electrodes and because of luminescence of vaporized electrode material and other constituents of the surrounding gaseous atmosphere. Considerable spread in the luminance, total radiation, and spectral power distribution may be achieved by varying the electrode materials.

LUMINESCENCE 16-20 Radiation from luminescent sources results from the excitation of single valence electrons of an atom, either in a gaseous state, where each atom is free from interference from its neighbors, or in a crystalline solid or organic molecule, where the action of its neighbors exerts a marked effect. In the first case, line spectra result, such as those of mercury or sodium arcs. In the second case, narrow emission bands result, which cover a portion of the spectrum (usually in the visible region). Both cases contrast with the radiation from incandescent sources, where the irregular excitation at high temperature of the free electrons of innumerable atoms gives rise to all wavelengths of radiation to form a continuous spectrum of radiation, as discussed in "Blackbody Radiation" above.

Photoluminescence Gaseous Discharge. Radiation, including light, can be produced by gaseous discharges as discussed previously under "Atomic Structure and Radiation." A typical mechanism for generating light (photons) from a gaseous discharge (such as in a fluorescent lamp) is described below (Figure 1-8). 1. A free electron emitted from the cathode collides with one of the two valence electrons of a mercury atom and excites it by imparting to it part of the kinetic energy of the moving electron, thus raising the valence electron from its normal energy level to a higher one.

Figure 1-8. Simplified energy diagram for mercury, showing a few of the characteristic spectral lines. 2. The conduction electron loses speed in the impact and changes direction, but continues along the tube to excite or ionize one or more additional atoms before losing its energy stepwise and completing its path. It generally ends at the wall of the tube, where it recombines with an ionized atom. A part of the electron current is collected at the anode. 3. Conduction electrons, either from the cathode or formed by collision processes, gain energy from the electric field, thus maintaining the discharge along the length of the tube.

4. After a short delay the valence electron returns to its normal energy level, either in a single transition or by a series of steps from one excited level to a lower level. At each of these steps a photon (quantum of radiant energy) is emitted. If the electron returns to its normal energy level in a single transition, the emitted radiation is called resonance radiation (Figure 1-9). 5. In some cases (as in the high-pressure sodium lamp) a portion of the resonance radiation is self-absorbed by the gas of the discharge before it leaves the discharge envelope. The absorbed energy is then re-radiated as a continuum on either side of the resonant wavelength, leaving a depressed or dark region at that point in the spectrum. Fluorescence. In the fluorescent lamp, UV radiation resulting from luminescence of the mercury vapor due to a gas discharge is converted into light by a phosphor coating on the inside of the tube or outer jacket. If this emission continues only during the excitation, it is called fluorescence. Figure 1-9 shows schematically a greatly magnified section of a part of a fluorescent lamp.

Figure 1-9. Magnified cross section of a fluorescent lamp, schematically showing progressive steps in the luminescent process, which finally result in in the release of visible radiation.

Figure 1-10. Fluorescence curve of a typical phosphor, showing initial excitation by ultraviolet rays and subsequent release of visible radiation.

Figure 1-11. Simplified energy diagram for a typical phosphor.

Figure 1-12. Color Characteristics of Important Fluorescent Lamp Phosphors The phosphors used in fluorescent lamps are crystalline inorganic compounds of exceptionally high chemical purity and of controlled composition to which small quantities of other substances (the activators) have been added to convert them into efficient fluorescent materials. With the right combination of activators and inorganic compounds, the color of the emission can be controlled. A typical schematic model for a phosphor is given in Figure 1-10, and an energy diagram for a typical phosphor is shown in Figure 1-11. In the normal state the electron oscillates about position A on the energy curve in Figure 1-11, as the lattice expands and contracts due to thermal vibration. For the phosphor to emit light it must first absorb radiation. In the fluorescent lamp this is chiefly at 253.7 nm. The absorbed energy transfers the electron to an excited state at position B. After loss of excess energy to the lattice as vibrational energy (heat), the electron again oscillates around a stable position C for a very short time, after which it returns to position D on the normal energy curve, with simultaneous emission of a photon of radiation. Stokes' law, stating that the radiation emitted must be of longer wavelength than that absorbed, is readily explained by this model. It then returns to A with a further loss of energy as heat and is ready for another cycle of excitation and emission. Because of the oscillation around both stable positions A and C, the excitation and emission processes cover ranges of wavelength, commonly referred to as bands.

In some phosphors two activators are present. One of these, the primary activator, determines the absorption characteristics and can be used alone, as it also gives emission. The other, the secondary activator, does not enter into the absorption mechanism but receives its energy by transfer within the crystal from a neighboring primary activator. The emitted light from the secondary activator is longer in wavelength than that from the primary activator. The relative amount of emission from the two activators is determined by the concentration of the secondary activator. The phosphors now used in most "white" fluorescent lamps are doubly activated calcium halophosphate phosphors in combination with rare-earth-activated phosphors. Figure 1-12 shows the characteristic colors and uses of phosphors currently employed in the manufacture of fluorescent lamps. Figure 1-13 gives the characteristics of some phosphors useful with mercury and metal halide lamps. Impurities other than activators and excessive amounts of activators have a serious deleterious effect on the efficiency of a phosphor.20 Phosphorescence. In some fluorescent materials, electrons can be trapped in metastable excited states for a time varying from milliseconds to days. After release from these states they emit light. This phenomenon is called phosphorescence. The metastable states lie slightly below the usual excited states responsible for fluorescence, and energy usually derived from heat is required to transfer the electron from the metastable state to the emitting state. Since the same emitting state is usually involved, the color of fluorescence and phosphorescence is generally the same for a given phosphor. In doubly activated phosphors the secondary activator phosphoresces longer than the primary activator, so the color changes with time. Short-duration phosphorescence is important in fluorescent lamps in reducing flicker in alternating current (ac) operation. Phosphors activated by IR radiation have an unusual type of phosphorescence. After excitation they show phosphorescence, which becomes invisible in a few seconds. However, they retain a considerable amount of energy trapped in metastable states, which can be released as light by IR radiation of the proper wavelength. Solid Laser.21-23 Lasers (light amplification by stimulated emission of radiation) are of major interest to illuminating engineers (see Chapter 6, Light Sources). In addition to amplifying light, lasers produce intense, highly monochromatic, well-collimated, coherent light.

Figure 1-13. Color Characteristics of Some Phosphors for Mercury and Metal Halide Lamps Coherent light consists of radiation whose waves are in phase with regard to time and space. Ordinary light, although it may contain a finite proportion of coherent light, is incoherent because the atomic processes that cause its emission occur in a random fashion. In a laser, however, electronic transitions are triggered (stimulated) by a wave of the same frequency as the emitted light. As a consequence, a beam of light is emitted, all of whose waves are in phase and of the same frequency. A prerequisite to laser action is a pumping process whereby an upper and a lower electron level in the active material undergo a population inversion. The pumping source may be a light, as in a ruby laser, or electronic excitation, as in a gas laser. The choice of laser materials is quite limited. First, it must be possible to highly populate an upper electronic level; second, there must be a light-emitting transition from this upper level with a long lifetime; third, a lower level must exist that can be depopulated either spontaneously or through pumping. Laser construction is as important to laser action, as is the source material. Since light wavelengths are too short to allow building a resonant cavity, long multi-nodal chambers are made with parallel reflectors at each end to feed back radiation until lasing takes place. The effect is to produce well-collimated light that is highly directional. Consider as an example the pink ruby laser, whose electronic transitions are shown in Figure 1-14, and whose mechanical

construction is indicated in Figure 1-15. This laser is pumped by a flash tube (a), and electrons in the ruby (b) are raised from level E1 to E3. The electrons decay rapidly and spontaneously from E3 to E2. They can then spontaneously move from E2 to E1 and slowly emit fluorescent light, hν21 (see Equation 1-3), or they can be stimulated to emit coherent light, hν21. The full reflector (c) and the partial reflector (d) channel the coherent radiation, hν21, until it has built up enough to emit coherent light hν21 through (d). The fact that this light has been reflected many times by parallel mirrors ensures that it is well collimated. The electrons are then available for further pumping (Figure 1-16).

Figure 1-14. Simplified diagrammatic representation of electronic transitions in a ruby laser. Gas Laser. In a solid laser there are three requirements: a material that reacts energetically to light, a population inversion generated by pumping in energy at the correct level and a growth of the internal energy caused by the reflection of photons within the solid. While the same requirements are met in a gas laser, two other characteristics are available, namely strong, narrow spectral lines and unequal emission at different energy levels. An example of such a gas laser is that containing a mixture of helium and neon (Figure 1-17). Helium is used as the energizing gas because it has a level from which it can lose energy only by collision. This level corresponds to the one at which neon radiates energy in the form of red light. On energizing helium in a gas discharge inside a cavity whose ends are reflecting and that contains both helium and neon, the helium transfers energy by collision with neon. The excited neon emits photons, which begin to amplify by cascading between the two reflecting surfaces until the internal energy is so large that the losses through the partially transmitting mirror become equal to the internal gains and the laser becomes saturated.

Figure 1-15. Simplified diagram of a ruby laser.

Figure 1-16. Photon cascade in a solid laser. Before the buildup begins, atoms in the laser crystal are in the ground state (a). Pumping light [arrows in (b)] raises most of the atoms to the excited state. The cascade (c) begins when an excited atom spontaneously emits a photon parallel to the axis of the crystal (photons emitted in other directions pass out of the crystal). The buildup continues in (d) and (e) through thousands of reflections back and forth from the silvered surfaces at the ends of the crystal). When amplification is great enough, light passes out at (f).

Figure 1-17. Structure of helium-neon gas laser, showing essential parts. Operation of the laser depends on the right mixture of helium and neon to provide an active medium. A radio-frequency exciter puts energy into the medium. The output beam is built up by repeated passes back and forth between reflecting end plates. Semiconductor Laser. A third type of laser uses a semiconducting solid material where the electron current flowing across a junction between p-type (electron-deficient) and n-type (electron-rich) material produces extra electrons in the conduction band (Figure 1-18). These radiate upon their transition back to the valence band or lower-energy states. If the junction current is large enough, there will be more electrons near the edge of the conduction band than there are at the edge of the valence band, and a population inversion may occur. To use this effect, the semiconductor crystal is polished with two parallel faces perpendicular to the junction plane. The amplified waves can then propagate along the plane of the junction and are reflected back and forth at the surfaces.

Electroluminescence24 Certain phosphors convert ac energy directly into light, without using an intermediate step as in a gas discharge, by utilizing the phenomenon of electroluminescence.

Figure 1-18. Diagram of an LED p-n junction.

Figure 1-19. Diagrammatic cross section of an electroluminescent lamp. Electroluminescent Lamps (ac capacitive). An electroluminescent lamp is composed of a two-dimensional area conductor (transparent or opaque) on which a dielectric-phosphor layer is deposited. A second two-dimensional area conductor of transparent material is deposited over the dielectric-phosphor mixture. An alternating electric field is established between the two conductors with the application of a voltage across the twodimensional (area) conductors. Under the influence of this field, some electrons in the electroluminescent phosphor are excited. During the return of these electrons to their ground or normal state the excess energy is radiated as light. Figure 1-19 shows a cross-sectional view of an electroluminescent lamp. Figure 1-20 gives the properties of some electroluminescent phosphors. The color of the light emitted by an electroluminescent lamp is dependent on frequency, while the luminance is dependent on frequency and voltage. These effects vary from phosphor to phosphor. The efficacy of electroluminescent devices is low compared to incandescent lamps. It is of the order of a few lumens per watt. Light-Emitting Diodes. Light-emitting diodes (LEDs) produce light by electroluminescence when low-voltage direct current is applied to a suitably doped crystal containing a p-n junction (Figure 1-18). The doping is typically carried out with elements from column III and V of the periodic table of elements. When activated by a forward biased current, If, the p-n junction emits light at a wavelength defined by the active region energy gap, Eg. The phenomenon was observed as early as 1923 in naturally occurring junctions, but was not considered practical due to its low luminous efficacy in converting electric energy to light. Efficacy has increased considerably since then such that LEDs are used for signals, indicators, signs, and displays.

Figure 1-20. Properties of Some Electroluminescent Phosphors When the forward biased current If is applied, minority carrier electrons are injected into the p-region and corresponding minority carrier electrons are injected into the n-region. Photon emission occurs as a result of electron-hole recombination in the p-region. Electron energy transitions across the energy gap, called radiative recombinations, produce photons (i.e., light), while shunt energy transitions, called nonradiative recombinations, produce phonons (i.e., heat). The energy band gap Eg, shown in Figure 1-18, is the separation between the conduction energy band and the valence energy band in the semiconductor crystal. The characteristics of the energy band gap determine the quantum efficiency and the radiative wavelengths of the LED device. For example, the radiative energy wavelength, λ, is given by

where h is Planck's constant and c is the speed of light. The luminous efficacies of typical AlInGaP LEDs and InGaN LEDs for different peak wavelengths are shown in Figure 1-21. The efficacy is dependent on the visible energy generated at the junction and losses due to reabsorption when light tries to escape through the crystal. Due to the high index of refraction of most semiconductors, light is reflected back from the surface into the crystal and highly attenuated before finally exiting. The efficacy expressed in terms of this ultimate measurable visible energy is called the external efficacy. The external efficacies are moderate, though the internal efficacies are calculated to be very high. For more information see Chapter 6, Light Sources. Cathodoluminescence. Cathodoluminescence is light emitted when a substance is bombarded by an electron beam from a cathode, as in cathode-ray and television picture tubes.

Figure 1-21. Properties of AlInGaP and InGaN LEDs

Miscellaneous Luminescence Phenomena Galvanoluminescence. Galvanoluminescence is light that appears at either the anode or the cathode when solutions are electrolyzed. Crystalloluminescence. Crystalloluminescence (lyoluminescence) is observed when solutions crystallize; it is believed to be due to rapid reformation of molecules from ions. The intensity increases upon stirring, perhaps on account of

triboluminescence (see below). Chemiluminescence. Chemiluminescence (oxyluminescence) is the production of light during a chemical reaction at room temperatures. True chemiluminescences are oxidation reactions involving valence changes. Thermoluminescence. Thermoluminescence is luminescence exhibited by some materials when slightly heated. In all cases of thermoluminescence, the effect is dependent on some previous illumination or radiation of the crystal. Diamonds, marble apatite, quartz, and fluorspar are thermoluminescent. Triboluminescence. Triboluminescence (piezoluminescence) is light produced by shaking, rubbing, or crushing crystals. Triboluminescent light may result from unstable light centers previously exposed to some source or radiation, such as light, X rays, radium emissions, and cathode rays; centers not exposed to previous radiation but characteristic of the crystal itself; or electrical discharges from fracturing crystals. Sonoluminescence. Sonoluminescence is light that is observed when sound waves are passed through fluids. It occurs when the fluids are completely shielded from an electrical field and is always connected with cavitation (the formation of gas or vapor cavities in a liquid). It is believed the minute gas bubbles of cavitated gas develop a considerable charge as their surface increases. When they collapse, their capacitance decreases and their voltage rises until a discharge takes place in the gas, causing a faint luminescence. Radioluminescence. Radioluminescence is light emitted from a material under bombardment from α rays, β rays, δ rays, or X rays.

LIGHT DETECTION Historically, the eye was used for most photometric assessments. Today, physical detectors have all but eliminated visual assessment for photometric purposes. Two common physical detector types in use today are photodiodes and photomultiplier tubes. Thermal detectors and photoconductive detectors are used for IR measurements.

Photodiodes Photodiodes are the most commonly used photodetectors for photometry and radiometry. Because of their excellent linearity and stability (freedom from fatigue), they replaced selenium cells, which had been widely used. Photodiodes are based on solid-state p-n junctions that react to external stimuli such as light. Rather than emitting light for the LED p-n junction, photons are absorbed in the p-n junction (Figure 1-18). Detectors are made of specific solid-state materials such as silicon, germanium, and indium-gallium-arsenide (InGaAs). Silicon photodiodes have sensitivity from the UV to nearIR region of the spectrum, and their spectral responsivity generally increases approximately linearly with wavelength throughout the visible region of the spectrum. Combined with a filter for photopic spectral response, silicon photodiodes are commonly employed in photometers. Recent high-quality silicon photodiodes have a dynamic range of eight orders of magnitude or larger and can also be used with special electronics for very low levels where photomultipliers had been required. Based on the quantum physics of photodiodes, some types of high-quality silicon photodiodes can be used as highaccuracy radiometric standards. This method, called the silicon photodiode self-calibration technique, was introduced during the late 1970s.25,26 Today, the highest-accuracy radiometric standards employ cryogenic radiometers, but silicon photodiodes are widely used as the most stable transfer standards in radiometry in the visible and near-IR region of the electromagnetic spectrum.

Photomultiplier Tubes Photomultiplier tubes (PMTs) are widely used as detectors for photometric and radiometric applications requiring high sensitivity (Figure 1-22). A PMT is a vacuum tube with a photocathode, a number of dynodes (i.e., a series of electrodes), and an anode. High voltages are applied between photocathode and dynodes and anode. The first element, the photocathode, is negatively biased and will eject photons (called photoelectrons) in response to radiant energy, due to the photoelectric effect. The photoelectrons hit the next dynodes with higher energy, creating more electrons (secondary electrons), which flow to the next dynode where even more electrons are emitted, eventually causing a cascade effect that multiplies the original number of photoelectrons by several orders of magnitude. Thus, photomultipliers have very high sensitivity. Spectral response ranges depend on the photocathode and the type of glass in the outer envelope, but they generally cover the visible region. Some others extend to the UV and near-IR regions of the spectrum. The stability of the voltage supply to PMT is especially critical to accurate measurements. Silicon photodiodes generally are more stable than PMTs. Photometers employing a PMT generally require an internal calibration source.

Figure 1-22. Schematic diagram of a photomultiplier and its electric circuit. From: G. Wyszecki, and W. Stiles, Color science. Copyright © 1982. Reprinted with permission of John Wiley & Sons, Inc.

Thermal Detectors Thermopiles and bolometers are known as thermal detectors. Thermal detectors have a light-receiving surface coated with black material such as carbon black and gold black. When light is incident on the black surface, it causes the surface temperature to rise due to absorbed radiation. The increase in temperature is proportional to the power of the absorbed radiation. Thermopiles employ a series of thermocouples to measure temperature. Bolometers employ metal or semiconductor materials having temperature-dependent resistance. Thermal detectors are seldom used in photometry due to their low sensitivity (several orders of magnitude lower than silicon photodiodes) and slow response time. An advantage of thermal detectors, however, is that they have generally nonselective spectral response, and are consequently well suited for radiant power measurements. Thermal detectors are often used in the IR region of the spectrum where other quantum detectors are not available.

Photoconductive Detectors Photoconductive detectors are semiconductors whose resistance changes directly as a result of photon absorption. These detectors use materials such as lead sulfide (PbS), lead selenide (PbSe), mercury cadmium telluride (HgCdTe), cadmium sulfide (CdS), and cadmium selenide (CdSe). Photoconductive detectors are widely used for IR measurements.

OPTICAL CONTROL27-29 Optical control may be provided in a number of ways. All are applications of one or more of the following phenomena: reflection, refraction, polarization, interference, diffraction, diffusion, and absorption.

Figure 1-23. The law of reflection states that the angle of incidence, θi, equals the angle of reflection, θr.

Reflection and Reflectors Reflection is the process by which a part of the light falling on a medium leaves that medium from the incident side. Reflection may be specular, spread, diffuse or compound, and selective or nonselective. Reflection from the front of a transparent plate is called first-surface reflection, and that from the back is called second-surface reflection. Refraction

and absorption by supporting media are avoided in first-surface reflection. Specular Reflection. If a surface is polished, it reflects specularly; that is, the angle between the reflected ray and the normal to the surface will equal the angle between the incident ray and the normal, as shown in Figure 1-23. If two or more rays are reflected, they may produce a virtual, erect, or inverted image of the source. Specular Reflectors. Examples of specular reflectors are: 1. Smooth polished metal and aluminized or silvered smooth glass or plastic surfaces. Reflector lamps use first-surface reflection when the bulb interior is coated with a thin metal reflecting mirror surface, as shown in Figure 1-24b. Light reflected from the upper surface of a transparent medium, such as glass plate, as in Figure 1-24a and c, also is an example of first-surface reflection. As shown in Figure 1-25, less than 5% of the incident light is reflected at the first surface unless it strikes the surface at wide angles from the normal. The sheen of silk and the shine from smooth or coated paper are images of light sources reflected in the first surface. 2. Rear-surface mirrors. Some light, the quantity depending on the incident angle, is reflected by the first surface. The rest goes through the transparent medium to a rear-surface mirror coating, where it is reflected as shown in Figure 1-24c. Reflection from Curved Surfaces. Figure 1-26 shows the reflection of a beam of light by a concave surface and by a convex surface. A ray of light striking the surface at point T obeys the law of reflection (Figure 1-25), and by taking each ray separately, the paths of various reflected rays may be constructed.

Figure 1-24. Reflections from (a) a transparent medium, such as clear plate glass, and from (b) front-surface and (c) rear-surface mirrors. In the case of parallel rays reflected from a concave surface, all the rays can be directed through a common point F by properly designing the curvature of the surface. This is called the focal point. The focal length FA is denoted by f. Spread Reflection. If a reflecting surface is not smooth (that is, corrugated, etched, or hammered), it spreads parallel rays into a cone of reflected rays, as shown in Figure 1-27b. Spread Reflectors. Slightly textured or hammered surfaces reflect individual rays at slightly different angles, but all in the same general direction. These are used to smooth beam irregularities and where moderate control or minimum beam spread is desired.

Figure 1-25. Effect of angle of incidence and state of polarization on the percentage of light reflected at an airglass surface: (a) Light that is polarized in the plane of incidence; (b) unpolarized light; (c) light that is polarized in a plane perpendicular to the plane of incidence.

Figure 1-26. Focal point and focal length of curved surfaces. Corrugated, brushed, dimpled, etched, or pebbled surfaces consist of small specular surfaces in irregular planes. Brushing the surface spreads the image at right angles to the brushing. Pebbled, peened, or etched surfaces produce a random patch of highlights. These are used where wide beams free from striations and filament images are required. The angle through which reflections are spread can be controlled by proper peening, for which equations describing peen radius and depth are available. Shot- or sandblasting and etching may cause serious losses in efficiency as a result of multiple reflections in random directions. Diffuse Reflection. If a material has a rough surface or is composed of minute crystals or pigment particles, the reflection is diffuse. Each ray falling on an infinitesimal particle obeys the law of reflection, but as the surfaces of the particle are in different planes, they reflect the light at many angles, as shown in Figure 1-27c. Diffuse Reflectors. Flat paints and other matte finishes and materials reflect at all angles and exhibit little directional control. These are used where wide distribution of light is desired.

Figure 1-27. The type of reflection depends on the surface: (a) polished surface (specular); (b) rough surface (spread); (c) matte surface (diffuse). Compound Reflection. Most common materials are compound reflectors and exhibit all three reflection components (specular, spread, and diffuse) to varying degrees. In some, one or two components predominate, as shown in Figure 128. Specular and narrowly spread reflections (usually surface reflections) cause the sheen on etched aluminum and semigloss paint. Diffuse-Specular Reflectors. Porcelain enamel, glossy synthetic finishes, and other surfaces with a shiny transparent finish over a matte base exhibit no directional control except for a specularly reflected ray as shown in Figure 1-28a, with an intensity of approximately 5 to 15% of the incident light.

Figure 1-28. Examples of compound reflection: (a) diffuse and specular; (b) diffuse and spread; (c) specular and spread.

Figure 1-29. Total reflection occurs when sin r = 1. The critical angle ic varies with the medium. Total Reflection. Total reflection of a light ray at a surface of a transmitting medium (Figure 1-29) occurs when the angle of incidence (θi) exceeds a certain value whose sine equals n2/n1, the ratio of indices of refraction. If the index of refraction of the first medium (n1) is greater than that of the second medium (n2), sin θr will become unity when sin θi is equal to n2/n1. At angles of incidence greater than this critical angle, the incident rays are reflected totally (Figure 1-30). In most glass total reflection occurs whenever sin θi is greater than 0.66, that is, for all angles of incidence greater than 41.8° (glass to air). Light piping by edge lighting and light transmission through rods and tubes are examples of total (internal) reflection. When light, passing through air, strikes a piece of ordinary glass (n2/n1 ≈ 1.5) normal to its surface, approximately 4% is reflected from the upper surface and 4% from the lower surface. Approximately 92% of the light is transmitted. The proportion of reflected light increases as the angle of incidence is increased (Figure 1-25).

Figure 1-30. Representation of light transmission through a single fiber of a fiber-optics system, showing (a) internal reflections and (b) the effect of light source location on collimation of light. Fiber Optics. Fiber optics is the branch of optical science concerned with thin, cylindrical glass or plastic fibers of optical quality. Light entering one end of the fiber is transmitted to the other end through the process of total internal reflection (Figure 1-30). In order to prevent light leaking from a fiber, it is coated with a lower-refractive-index material. Large numbers of fibers (from 100 to 1,000,000) can be clustered together to form a bundle. Fiber bundles are of two major types: coherent and noncoherent. The first are used for transmitting images, and each individual fiber is carefully oriented with respect to its neighbors in the entire bundle. Noncoherent bundles have random fiber locations in the bundle, but are suitable for transmitting light between points.

Refraction and Refractors A change in the velocity of light (speed of propagation, not frequency) occurs when a ray leaves one material and enters another of greater or lower optical density. The speed will be reduced if the medium entered is denser, and increased if it is less dense. Except when light enters at an angle normal to the surface of the new medium, the change in speed always is accompanied by a bending of the light from its original path at the point of entrance, as shown in Figure 1-31. This is known as refraction. The degree of bending depends on the relative densities of the two substances, on the wavelength of the light, and on the angle of incidence, being greater for large differences in density than for small. The light is bent toward the normal to the surface when it enters a denser medium, and away from the normal when it enters a less dense

material. When light is transmitted from one medium to another, each ray follows the law of refraction. When rays strike or enter a new medium, they may also be scattered in many directions because of irregularities of the surface, such as fine cracks, mold marks, scratches or changes in contour, or because of foreign deposits of dirt, grease, or moisture.

Figure 1-31. Refraction of light rays at a plane surface causes bending of the incident rays and displacement of the emergent rays. A ray passing from a rare to a denser medium is bent toward the normal to the interface, while a ray passing from a dense to a rarer medium is bent away from the normal. Snell's Law. The law of refraction (Snell's law) is expressed as follows:

where n1 = index of refraction of the first medium, θi = angle the incident light ray forms with the normal to the surface, n2 = index of refraction of the second medium, θr = angle the refracted light ray forms with the normal to the surface. When the first medium is air, of which the index of refraction usually is taken as 1 (the vacuum value; this approximation is correct to three decimal places), the formula becomes

The two interfaces of the glass plate shown in Figure 1-31 are parallel, and therefore the entering and emerging rays also are parallel. The rays are displaced from each other (a distance D) because of refraction. Examples of Refraction. A common example of refraction is the apparent bending of a straw at the point where it enters the water in a drinking glass. Although the straw is straight, light rays coming from that part of the straw under water are refracted when they pass from the water into the air and appear to come from higher points. Prismatic light directors, such as shown in Figure 1-32a and b, may be designed to provide a variety of light distributions using the principles of refraction. Lens systems controlling light by refraction are used in automobile headlights and in beacon, floodlight, and spotlight Fresnel lenses. Prisms. Consider Snell's law:

This equation suggests, since the velocity of light is a function of the indices of refraction of the media involved and also of wavelength, that the exit path from a prism will be different for each wavelength of incident light and for each angle of

incidence (Figure 1-33). This orderly separation of incident light into its spectrum of component wavelengths is called dispersion.

Figure 1-32. Optical systems utilizing the refractive properties of prisms and lenses: (a) Street lighting unit in which the outer piece controls the light in vertical directions (concentrating the rays into a narrow beam at about 75° from the vertical) and the inner piece re directs the light in the horizontal plane. The result is a "twoway" type of intensity distribution. (b) Prismatic lens for a fluorescent lamp luminaire intercepts as much light as possible, redirecting part from the glare zone to more useful directions (c) Cylindrical and flat Fresnel lenses. (d) Reflecting prism. Refracting Prisms. The degree of bending of light at each prism surface is a function of the refractive indices of the media and the prism angle (A in Figure 1-33). Light can be directed accurately within certain angles by having the proper angle between the prism faces.

Figure 1-33. White light is dispersed into its component colors by refraction when passed through a prism. The angle of deviation D (illustrated for green light) varies with wavelength. Refracting prisms are used in such devices as headlight lenses and refracting luminaires. In the design of refracting equipment, the same general considerations of proper flux distribution hold true as for the design of reflectors. Following Snell's law of refraction, the prism angles can be computed to provide the proper deviation of the light rays from the source. For most commercially available transparent materials like glasses and plastics, the index of refraction lies between 1.4 and 1.6. Often, by proper placement of the prisms, it is possible to limit the prismatic structure to one surface of the refractor, leaving the other surfaces smooth for easy maintenance. The number and the sizes of prisms used are governed by several considerations. Among them are ease of manufacture and convenient maintenance of lighting equipment in service. Use of a large number of small prisms may magnify the effect of rounding of prisms that occurs in manufacture; on the other hand, small prisms produce greater accuracy of light control. Ribbed and Prismed Surfaces. These can be designed to spread rays in one plane or scatter them in all directions. Such surfaces are used in lenses, luminous elements, glass blocks, windows, and skylights. Reflecting Prisms. These reflect light internally, as shown in Figure 1-32d, and are used in luminaires and retro-directive markers. Their performance quality depends on the flatness of reflecting surfaces, accuracy of prism angles, elimination of dirt in optical contact with the surface, and elimination (in manufacturing) of prismatic error. Lenses. Positive lenses form convergent beams and real inverted images as in Figure 1-34a. Negative lenses form divergent beams and virtual, inverted images as in Figure 1-34b. Stepped and Fresnel Lenses. The weight and cost of glass in large lenses used in illumination equipment can be reduced by making cylindrical steps in the flat surface. The hollow, stepped back surface reduces the total quantity of glass used in the lens. In a method developed by Fresnel, as shown in Figure 1-33c, the curved face of the stepped lens becomes curved rings and the back is flat. Both the stepped and Fresnel lenses reduce the lens thickness, and the optical action is approximately the same. Although outside prisms are slightly more efficient, they are likely to collect more dust. Therefore, prismatic faces are often formed on the inside.

Figure 1-34. Ray path races through lenses: (a) positive; (b) negative.

Figure 1-35. Lens aberrations. (a) Spherical aberration: convergence of parallel rays at different focal points at different distances from the axis of a lens. (b) Coma: difference in the lateral magnification of rays passing through different zones of a lens. (c) Chromatism: a difference in focal length for rays of different wavelengths. (d) Astigmatism and curvature: existence in two parallel planes of two mutually perpendicular line foci and a curved image plane. (e) Distortion: a difference in the magnification of rays passing through a lens at different angles. Lens Aberrations. There are, in all, seven principal lens aberrations: spherical aberration, coma, axial and lateral chromatism, astigmatism, curvature, and distortion (Figure 1-35). Usually they are of little importance in lenses used in common types of lighting equipment. The simpler the lens system, the more difficult it is to correct the aberrations.

Transmission and Transmitting Materials Transmission is a characteristic of many materials: glass, plastics, textiles, crystals, and so forth. The luminous transmittance τ of a material is the ratio of the total emitted light to the total incident light; it is affected by reflections at each surface of the material, as explained in Figure 1-24, and by absorption within the material. Figure 1-36 lists characteristics of several materials. Bouguer's or Lambert's Law. Absorption in a clear transmitting medium is an exponential function of the thickness of the medium traversed:

where I = intensity of transmitted light, I0 = intensity of light entering the medium after surface reflection, α = absorption coefficient that characterizes the absorbing properties of a unit thickness of the medium, τ = transmittance of a unit thickness, d = thickness of the medium traversed. The optical density D is the common logarithm of the reciprocal of the transmittance:

Spread Transmission. Spread transmission materials offer a wide range of textures. They are used for brightness control, as in frosted lamp bulbs, in luminous elements where accents of brilliance and sparkle are desired, and in moderately uniform brightness luminaire-enclosing globes. Care should be used in placing lamps to avoid glare and spotty appearance. Figure 1-37a shows a beam of light striking the smooth side of a piece of etched glass. In Figure 1-37b, the frosted side is toward the source, a condition that with many ground or otherwise roughened glasses results in appreciably higher transmittance. For outdoor use, the rough surface usually must be enclosed to avoid excessive dirt collection. Diffuse Transmission. Diffusing materials scatter light in all directions, as shown in Figure 1-37c. White, opal, and prismatic plastics and glass are widely used where uniform brightness is desired. Mixed Transmission. Mixed transmission is a result of a spectrally selective diffusion characteristic exhibited by certain materials such as fine opal glass, which permits the regular transmission of certain colors (wavelengths) while diffusing other wavelengths. This characteristic in glass varies greatly, depending on such factors as its heat treatment, composition, thickness, and the wavelengths of the incident light.

Figure 1-36. Reflecting and Transmitting Materials

Figure 1-37. (a) Spread transmission of light incident on the smooth surface of figured, etched, ground, and hammered glass samples. (b) Spread transmission of light incident on the rough surface of the same samples. (c) Diffuse transmission of light incident on solid opal and on flashed opal glass, white plastic or marble sheet. (d) Mixed transmission through opalescent glass.

Polarization Unpolarized light consists of visible electromagnetic waves having transverse vibrations of equal magnitude in an infinite number of planes, all of which oscillate about the line representing the direction of propagation (Figure 1-38). In explaining the properties of polarized light, it is common to resolve the amplitude of the vibrations of any light ray into components vibrating in two orthogonal planes each containing the light ray. These two principal directions are usually referred to as the horizontal and vertical vibrations. The horizontal component of light is the summation of the horizontal components of the infinite number of vibrations making up the light ray. When the horizontal and vertical components are equal, the light is unpolarized. When these two components are not equal, the light is partially or totally polarized as shown in Figure 1-38. The percentage polarization of light from a source or luminaire at a given angle is defined by the following relation:29

where Iv and Ih are the intensities of the vertical and horizontal components of light, respectively, at the given angle.

Figure 1-38. Graphical representations of polarized and unpolarized light. Reference to vertically polarized light or horizontally polarized light can be misleading in that it suggests that all light waves vibrate either horizontally or vertically. A better terminology would be to refer to light at a given instant as consisting of one component vibrating in a horizontal plane and another component vibrating in a vertical plane. A general terminology would identify the light components in terms of two reference planes as shown in Figure 1-39. One plane is the plane of the task at the point of the incident light ray, and the second plane is the plane of incidence: the plane perpendicular to the plane of the task and containing the incident light ray. Then the two components of light would be referred to as the parallel component, or the component in the plane of incidence, and the perpendicular component. This terminology would apply to any task position and would be free of ambiguity with respect to spatial orientation. Polarized light can be produced in four ways: (1) scattering, (2) birefringence, (3) absorption, and (4) reflection and refraction. Scattering is the mechanism of polarization in daylighting; that is, light from a clear blue sky is partially polarized due to

the scattering of light by particles in the air.

Figure 1-39. Reference planes of a task. The birefringence, or double refraction property, of certain crystals can be used to achieve polarization. However, the size of these crystals limits this technique to scientific applications; it is not suitable for general lighting. Polarization by absorption can be achieved by using dichroic polarizers. These polarizers absorb all of the light that is in one particular plane and transmit a high percentage of the light polarized in a perpendicular plane. A high percentage of polarization can be obtained by this method, but with a loss of total luminous transmittance. This type of polarizer is commonly used in sunglasses, where it is oriented to transmit the vertical component of light while suppressing the horizontal (typically reflected) component.

Figure 1-40. Polarization by reflection at a glass-air surface is at a maximum when the angle of incidence i plus the angle of refraction r equals 90°.

Figure 1-41. Principle of multilayer polarizers. Light may be polarized by utilizing the reflection characteristics of dielectric materials. When light is reflected from a glass surface, it is partially polarized; a larger percentage of the horizontal component is reflected than of the vertical component. At approximately 57° (Brewster's angle), the reflected light contains only the horizontal component (Figure

1-25). For this one surface, however, only 15% of the incident horizontal component is reflected. The light transmitted through a plate at this angle is made up of the remaining portion of the horizontal component and all the vertical component of the original beam. The resulting light is partially polarized (Figure 1-40). As additional glass plates are added to the system, more and more of the horizontal component is reflected and the transmitted light is more completely vertically polarized. A stack of glass plates, as shown in Figure 1-41, thus becomes a method of producing polarization, and the polarizing effect is greatest at Brewster's angle. The percentage polarization is less at all other angles and is zero for a light ray at normal incidence. Polarization by this method can be obtained by arranging glass or plastic flakes in a suitable material.

Interference When two light waves of the same wavelength come together at different phases of their vibration, they combine to make up a single wave whose amplitude is between the difference and the sum of the amplitudes of the two, depending on their relative phase. Figure 1-42 illustrates this concept for waves of water in a pool. The waves tend to cancel each other at the node lines. Figure 1-43 shows the resulting interference when light refracts and reflects from thin films. Part of the incident light ab is first reflected as bc. Part is refracted as bd, which again reflects as de, and finally emerges as ef. If waves bc and ef have wavefronts of appreciable width, they will overlap and interfere. Optical interference coatings have been used for many years in cameras, projectors, and other optical instruments and can reduce reflection from transmitting surfaces, separate heat from light, transmit or reflect light according to color, increase reflections from reflectors, or perform other light control functions. Naturally occurring examples of interference are soap bubbles and oil slicks. Also, many birds, insects, and fish get their iridescent colors from interference films. The application of interference coatings can significantly increase the reflectance of reflectors and the transmittance of luminaire glass or plastic enclosures. Low-Reflectance Films. Dielectric optical interference films are applied to surfaces to reduce reflectance, increase transmittance, and consequently improve contrast relationships. Films that are one-quarter wavelength thick with an index of refraction between that of the medium surrounding the glass and that of the glass are used. The hardest and most permanent films are those of magnesium fluoride condensed on the transmitting surface after thermal evaporation in vacuum.

Figure 1-42. Interference. The usual 4% reflection at uncoated air-to-glass surfaces may be reduced to less than 0.5% at each filmed surface at normal incidence, as a result of the canceling interference between the waves reflected at the air-to-film and film-to-glass surfaces. Dielectric coatings can be made very specific to one reflected wavelength or, by varying the layer's thickness or index of refraction, spread over a wide wavelength interval. Dichroic (Dielectric) Coating. A multilayer coating that selectively transmits or reflects portions of the spectrum can be added to optical materials. Often called hot or cold mirrors, such coatings are efficient in their selective reflection and transmission, respectively, of IR energy. The coatings are typically designed for incident radiation at 45° or 90° to the coated surface. Deviations from the design angle will change the reflected and the transmitted energy. Undesirable results occur when dichroic filters are used in wide beams of light, since the color varies across the resulting beam. Hot-mirror lamp envelopes, which reflect IR back to a filament, are used with special tungsten-halogen lamps to increase their efficacy without increasing their wattage and reducing their life.

Figure 1-43. Constructive and destructive interference.

Diffraction Due to its wave nature, light will be redirected as it passes by an opaque edge or through a small slit. The wavefront broadens as it passes by an obstruction, producing an indistinct, rather than sharp, shadow of the edge. The intensity and spatial extent of the shadow depends on the geometric characteristics of the edge, the physical extent (size and shape) of the source, and the spectral properties of the light. Light passing through a small slit will produce alternating light and dark bars as the wavefronts created by the two edges of the slit interfere with one another.

Diffusion Diffusion is the breaking up of a beam of light and the spreading of its rays in many directions by irregular reflection and refraction from microscopic crystalline particles, droplets or bubbles within a transmitting medium, or from microscopic irregularities of a reflecting surface. Perfect diffusion seldom is attained in practice but sometimes is assumed in calculations in order to simplify the mathematics (Figures 1-27c).

Absorption Absorption occurs when a light beam passes through a transparent or translucent medium or meets a dense body such as an opaque reflector surface. If the intensity of all wavelengths of the light passing through a transparent body is reduced by nearly the same amount, the substance is said to show general absorption. The absorption of certain wavelengths of light in preference to others is called selective absorption. Most colored objects owe their color to selective absorption in some part of the visible spectrum, with resulting reflection and transmission in other selected parts of the spectrum.

REFERENCES 1. Richtmyer, F. K., E. H. Kennard, and J. N. Cooper. 1969. Introduction to modern physics. 6th ed. New York: McGraw-Hill. 2. Born, M. 1989. Atomic physics. 8th rev. ed. NewYork: Dover Publishing. 3. Born, M., and E. Wolf. 1998. Principles of optics: Electromagnetic theory of propagation, interference and diffraction of light. 6th reissued ed. New York: Pergamon Press. 4. Elenbaas, W. 1972. Light sources. New York: Crane, Russak & Co. 5. Maxwell, C. J. 1954. A treatise on electricity and magnetism. 3rd ed. New York: Dover Publications. 6. Forsythe, W. E. 1937. Measurement of radiant energy. New York: McGraw-Hill. 7. Commission Internationale de l'Éclairage. 1994. CIE collection in photometry and radiometry. CIE no. 114. Vienna: Bureau Central de la CIE. 8. Goodeve, C. F. 1936. Relative luminosity in the extreme red. Proc. R. Soc. Lond. Ser. A 155(886):664-683. 9. Commission Internationale de l'Éclairage. 1978. Light as a true visual quantity: Principles of measurement, CIE no. 41. Vienna: Bureau Central de la CIE.

10. Commission Internationale de l'Éclairage. 1983. The basis of physical photometry, CIE Publication no. 18.2. Paris: Bureau Central de la CIE. 11. Gibson, K.S., and E. P. T. Tyndall. 1923. Visibility of radiant energy. Bulletin Bureau of Standards 19:131. 12. Commission Internationale de l'Éclairage. 1990. CIE 1988 2° spectral luminous efficiency function for photopic vision. CIE no. 86. Vienna: Bureau Central de la CIE. 13. He, Y., A. Bierman, and M. S. Rea. 1998. A system of mesopic photometry. Light. Res. Tech. 30(4):175-181. 14. He, Y., M. S. Rea, and J. Bullough. 1997. Evaluating light source efficacies under mesopic conditions using reaction times. J. Illum. Eng. Soc. 26(1):125-138. 15. Commission Internationale de l'Éclairage. 1989. Mesopic photometry: History, special problems and practical solutions. CIE no. 81. Vienna: Bureau Central de la CIE. 16. Waymouth, J. F. 1971. Electric discharge lamps. Cambridge: MIT Press. 17. Fonda, G. R., and F. Seitz, eds. 1948. Preparation and characteristics of solid luminescent materials. New York: John Wiley. 18. Leverenz, H. W. 1950. An introduction to luminescence of solids. New York: John Wiley. 19. Harvey, E. N. 1957. A history of luminescence from the earliest times until 1900. Philadelphia: American Philosophical Society. 20. Wachtel, A. 1958. The effect of impurities on the plaque brightness of a 3000° K calcium halophosphate phospher. J. Electrochem. Soc. 105(5):256-260. 21. Brotherton, M. 1964. Masers and lasers: How they work, what they do. New York: McGraw-Hill. 22. Harvey, A. F. 1970. Coherent light. London, New York: Wiley-Interscience. 23. Lengyel, B. A. 1966. Introduction to laser physics. New York: John Wiley. 24. Ivey, H. F. 1963. Electroluminescence and related effects. New York: Academic Press. 25. Geist, J. 1979. Quantum efficiency of the p-n junction in silicon as an absolute radiometric standard. Appl. Opt. 18(6): 760-762. 26. Zalewski, E. F., and J. Geist. 1980. Silicon photodiode absolute spectral response self-calibration. Appl. Opt. 19(8): 1214-1216. 27. IES. Committee on Light Control and Equipment Design. 1959. IES guide to design of light control. Part I: Physical principles. Part II: Design of reflector and optical elements. Illum. Eng. 54(2):722-786. 28. Resnick, R., and D. Halliday. 1977. Physics. 3rd. ed. New York: John Wiley. 29. Hardy, A. C., and F. H. Perrin. 1932. The principles of optics. New York: McGraw-Hill. 30. IES. Committee on Testing Procedures for Illuminating Characteristics. 1963. Resolution on reporting polarization. Illum. Eng. 58(5):386.

2 Measurement of Light and Other Radiant Energy PRINCIPLES OF PHOTOMETRY AND RADIOMETRY Introduction Progress in a branch of science or engineering is very much dependent on the ability to measure the associated quantities. Lord Kelvin (1824-1907) expressed this most bluntly: When you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely, in your thoughts, advanced to the stage of science, whatever the matter may be. The earliest instruments for measuring luminous quantities depended on visual appraisal. Such methods lacked both precision and accuracy, largely because the results were dependent on the individual observers making the measurement. Even for a particular observer, measurement reproducibility was poor because a number of variables influencing the measurements could not be controlled or explained. These visual methods are now rarely used. Today measurements usually are made using calibrated physical instruments that respond to radiant energy. The human visual system responds only to radiation in a very narrow band of the electromagnetic spectrum. This range of wavelengths is approximately from 380 to 780 nm, depending on the individual observer. It should be kept in mind that for any source of illumination, the radiant energy produced is rarely limited to wavelengths within these boundaries. Although the primary concern in this chapter is the measurement of radiation that results in visual sensation, measurements of radiant quantities outside the visible spectrum are also important because of the nonvisual effects that this radiation produces (see Chapter 5, Nonvisual Effects of Optical Radiation). Optical radiation generally refers to all radiation that can be measured using certain techniques and equipment (mirrors, lenses, filters, diffraction gratings, prisms). Thus visible, ultraviolet (UV), and infrared (IR) radiation are collectively considered as optical radiation. The measurement of optical radiation is called radiometry. Radiometry is the science of measuring radiant quantities without regard for the visual effects of the radiation. Light almost always refers to wavelengths visible to humans, although sometimes invisible radiation is also called light when describing radiation on plants or on skin. Photometry, a special branch of radiometry, is the measurement of radiation in terms of human visual response. The Commission Internationale de l'Éclairage (CIE) has established a standard observer response curve (also known as the photopic luminous efficiency function), denoted by V(λ) (see Chapter 1, Light and Optics). This standard observer response curve, with its peak at approximately 555 nm, is used as a standard weighting function that, when applied to a spectral power distribution (SPD) of the light being measured, is an approximation of the perceived brightness of that light. The standardization of the eye spectral sensitivity function is the key to photometry, removing the influence of the observer from the measurements. However, despite the industry-wide acceptance of this function, one should recognize that it represents a compromise in assuming a predictable correlation of physical measurements with visual response, and that there are some circumstances where the system works poorly (see Chapter 3, Vision and Perception).1 Photopic, Mesopic, and Scotopic Vision. Vision can be categorized with reference to the adaptive state of the rod and cone photoreceptors of the retina. At very low luminance levels, below approximately 0.01 cd/m2--the scotopic region--the light energy is insufficient to energize the cone photoreceptor system, but is adequate to stimulate the rod photoreceptor system. The standard luminous efficiency function for scotopic vision is represented by the function V′ (λ), with its peak near 507 nm (see Chapter 1, Light and Optics). At luminance levels greater than approximately 3 cd/m2--the photopic region--colors can be distinguished, and

objects having fine detail can be readily seen in the central visual field, where the density of the cone population of the retina is highest, that is, in the fovea. Strictly speaking, the photopic luminous efficiency function applies to visual fields of size 2° or less. At intermediate luminance levels, between approximately 0.01 and 3 cd/m2--the mesopic region--both rod and cone photoreceptors contribute to vision. Because of methodological difficulties, there is presently no standard luminous efficiency function for this range of adaptation luminance, although it is of practical importance for roadway, security, and other exterior nighttime conditions (see "mesopic vision" in Chapter 3, Vision and Perception).

Basic Concepts Units of Measurement. The International System of Units, abbreviated SI, is accepted worldwide as a standard system of units of measurement. In that system, the fundamental photometric quantity, luminous intensity, is expressed in candelas (cd). The magnitude of the candela has a historical basis. At one time called the candlepower, it was defined in terms of flame or filament standards. For practical purposes the terms candela and candlepower are equivalent and, although no longer standard, the latter term is still occasionally used. The current definition of the candela is2 . . . the luminous intensity, in a given direction, of a source that emits monochromatic radiation of frequency 540 × 1012 Hz and that has a radiant intensity in that direction of 1/683 W/sr. The candela is now formally defined on this radiometric basis because of advances that have been made in this area of metrology. The definition expresses the candela in terms of the watt and the steradian. The steradian is defined as the solid angle subtending an area on the surface of a sphere equal to the square of the sphere's radius. The steradian is an SI supplementary unit. The unit of power, the watt, is likewise not a base SI unit but can be defined as 1 J/s (energy per unit time) or, in base units, 1 m2 × kg × s−3. Since the formal definition of the candela is at a single wavelength (at the peak of the photopic luminous efficiency function), V(λ) must be applied to measurements of radiant power produced by real sources in order to reduce them to candelas. In casual discussions, the terms "energy" and "power" often are used interchangeably; in general discussions on measurements of optical radiation, the term "radiant energy" is most commonly used (as in the title of this chapter). However, it should be kept in mind that it is radiant power (that is, the transfer of energy per unit of time) weighted in terms of an eye sensitivity curve, and not radiant energy, that acts as the visual stimulus. The terms "radiant power" and "radiant flux" are used synonymously. There is no photopic weighting inherent in the concept of radiant flux; it is strictly a radiometric quantity. The unit of radiant flux is the watt. Two important derived units based on the candela are those of luminous flux and illuminance. To understand how these two quantities are related to the luminous intensity, consider a hypothetical model, illustrated in Figure 2-1. An isotropic point source of radiation (that is, one that radiates energy uniformly in all directions) is located at the geometric center of an ideal sphere of zero reflectance (all incident radiation is absorbed). Any portion of the inner sphere's surface receives only direct radiation from the point source itself and no reflected radiation from other parts of the sphere's surface. For a sphere having a radius of one unit, a one-square-unit area on the sphere's surface represents a solid angle of one steradian. (The two-dimensional shape of this area is irrelevant; it might be a circle, in which case the steradian would be represented by a cone. This becomes clearer when the strict definition of luminous intensity is given in terms of a limit.) The luminous intensity of the source in this model is the same in all directions and assumed to be 1 cd. The radiant flux falling on a unit area of the sphere's surface can now be defined to be the luminous flux.

Figure 2-1. Relationship between candelas, lumens, lux, and footcandles. A point source (luminous intensity = 1 cd) is shown at the center of a sphere of unit radius whose surface has a reflectance of zero. The illuminance at any point on the sphere is 1 Ix if the radius is 1 m, or 1 fc if the radius is 1 ft. The solid angle subtended by the area ABCD is 1 sr. The flux density is therefore 1 lm /sr, which corresponds to a luminous intensity of 1 cd as originally assumed. The sphere has a total area of 4π m2 or ft2, and there is a luminous flux of 1 lm falling on each unit area. Thus, the source provides a total of 4π lm. The unit of luminous flux is the lumen (lm); the quantity of luminous flux falling on one square unit of the sphere's surface is defined as 1 lm. Note that the unit itself is arbitrary, since the total quantity of flux that will be incident on this area is independent of the size of the sphere. For a sphere of unit radius, it can be shown by simple geometry that the area of the sphere's surface is equal to 4π square units; thus the isotropic source having a luminous intensity of 1 cd produces a total luminous flux of 4π lm. The concentration of luminous flux falling on a surface, that is, the incident flux per unit area, is called illuminance. To define a unit of illuminance, the sphere must now be given real dimensions because the flux density diminishes with increasing distance from the source. If the sphere's radius is 1 m, the illuminance on the sphere's wall is 1 lm/m2, or 1 lux (lx). If the radius is 1 ft, the illuminance is 1 lm/ft2, or 1 footcandle (fc). Another important luminous quantity is luminance. This quantity is more difficult to grasp, and the sphere model is not useful for that purpose. Luminance relates directly to perceived brightness, that is, the visual effect that illumination produces. Luminance depends not only on the illuminance on an object and its reflective properties, but also on its projected area on a plane perpendicular to the direction of view. There is a direct relationship between the luminance of a viewed object and the illuminance of the resulting image on the retina of the eye. This is analogous to the exposure requirements in photography. The unit of luminance is the candela per square meter (cd/m2). Photometric quantities, along with their radiometric counterparts, are discussed and defined in the Glossary. Radiometric and photometric measurements frequently involve a consideration of the inverse square law (which is strictly applicable only for point sources) and the cosine law. Inverse Square Law. The inverse square law (Figure 2-2a) states that the illumination E at a point on a surface varies directly with the luminous intensity I of the source, and inversely as the square of the distance d between the source and the point. If the surface at the point is normal to the direction of the incident light, the law may be expressed as follows:

This equation holds true within 1% when d is at least five times the maximum dimension of the source (or luminaire) as viewed from the point on the surface. For a further discussion of this "five-times rule," see Chapter 9, Lighting Calculations. Cosine Law. The cosine law (Figure 2-2b), also known as Lambert's law, states that the illuminance on any surface varies as the cosine of the angle of incidence. The angle of incidence, θ, is the angle between the normal to the surface and the direction of the incident light. The inverse square law and the cosine law can be combined as follows:

Cosine-Cubed Law. A useful extension of the cosine law is the cosine-cubed equation (Figure 2-2c). By substituting h/cosθ for d, the above equation may be written

Other Measurable Quantities. The principal photometric quantities have been discussed. These and other quantities of interest are summarized in Figure 2-3 and in the Glossary.

PHOTOMETRY IN PRACTICE: GENERAL REQUIREMENTS Traceability and Accreditation Contractual agreements often require that measurements made by laboratories be traceable to national and international measurement standards and that the laboratories be able to support claims of traceability with appropriate documentation and records of equipment used in the measurement process. Such agreements are becoming more common with increasing awareness and adoption of documented quality management systems based on the ISO 9000 series of standards, which states that: 4.11.2b [The supplier shall:] identify all inspection, measuring, and test equipment that can affect product quality, and calibrate and adjust them at prescribed intervals, or prior to use, against certified equipment having a known valid relationship to internationally or nationally recognized standards. Where no such standards exist, the basis for calibration shall be documented.

Figure 2-2. (a) The inverse-square law illustrating how the same quantity of light flux is distributed over a greater area, as the distance from source to surface is increased. (b) The Lambert cosine law showing that light flux striking a surface at angles other than normal is distributed over a greater area. (c) The cosinecubed law explaining the transformation of the formula. The requirement for traceability3 involves the ability to relate individual measurement results, with a stated uncertainty, through an unbroken chain of comparisons to a stated reference source. In this way, the reference source is transferred from the national standards through to the end user of the national measurement system.

Figure 2-3. Some Measurable Characteristics of Light, Light Sources, and Lighting Materials To assure the end user of traceability and competence throughout the chain, laboratories may seek recognition through assessment and accreditation by a recognized accreditation body. The assessments are discussed in Reference 49, which sets out the general management and technical requirements for calibration or testing laboratories. Laboratories that comply with this guide also comply, by definition, with the requirements of ISO 9002 for the scope of the testing and calibration services covered by their quality management system. The reverse is not true since ISO 9002 registration gives no recognition of the quality of the laboratory's specific measurements. In the United States, a number of federal and private agencies offer calibration laboratory assessment and accreditation services. Contact the National Institute of Standards and Technology (NIST) for additional information. In Canada, contact the National Research Council of Canada (NRC).

Standards4,5 Primary Standards. A primary standard is a standard that is designated or widely acknowledged as having the highest metrological quantities and whose value is accepted without reference to other standards of the same quantity. The candela, maintained by the Bureau International des Poids et Mesures (BIPM), is a primary standard. National (Measurement) Standards. National standards that define radiometric and photometric quantities are maintained by national standard laboratories.2 These standards typically are developed from international standards through a specified, usually complex, experimental procedure. In North America, measurement standards for lighting, including the candela, are maintained by NIST and NRC. National measurement standards are not directly accessible by other laboratories. Transfer Standards. Transfer standards are necessary to link the measurement systems of one laboratory (e.g., a national measurement laboratory) to another. They are defined simply as intermediaries used to compare standards. They can be called travelling standards when intended for transport between different locations. Reference Standards. Reference standards are standards having the highest metrological quantity available at a given location or in a given organization, from which the measurements made there are derived. Reference standards can be derived directly from a national measurement standard or from the reference standards of other laboratories in the calibration chain. They usually are prepared with precise electrical and radiometric measurement equipment. Working Standards. Working standards are used for routine measurements in a laboratory and usually are prepared and calibrated by that same laboratory from its own reference standard. The preceding classification of standards is based on the 1993 International Vocabulary of Basic and General Terms in Metrology published by the BIPM. Other nomenclatures have evolved from historical usage. They do not represent the internationally accepted BIPM definitions, and they are not all consistent. For example, the term "primary standard" often is used to designate a standard source that was obtained from a national standards laboratory and that is used only to make other working standards for everyday use in that laboratory. Sometimes, a primary standard is called a "master standard." The term "secondary standard" is also commonly used in private laboratories to distinguish a standard from the one called primary, and sometimes the terms "secondary standard" and "working standard" are used interchangeably. The term "tertiary standard" is used if there are three levels of standards

deployed. To avoid confusion, the BIPM definitions should be used. Documentation of Standards. It is most important to document the lineage of all standards back to the national measurement standards. The documentation should define the calibration procedure and also the state of any influencing environmental conditions. Measurement uncertainty should be documented in accordance with the ISO Guide to the Expression of Uncertainty in Measurement (1995). It should include the uncertainty due to random and systematic error. Without such documentation, measurements cannot be properly compared. Types of Standards. Incandescent lamps of various wattages are commonly used to establish a traceability chain for photometric quantities outside NIST and NRC. Alternative standards are calibrated detectors from which sources themselves can be calibrated. Such detectors can also be used to calibrate other photometric detectors. Ideal standard lamps and detectors have two principal characteristics: they both are accessible and invariant. In the United States, lamp and detector standards can be purchased from NIST. In Canada, consult NRC. Photometric standards can also be purchased from private laboratories. This adds one or more steps in the calibration chain and increases the uncertainties in the measured quantities. If the supplier of photometric standards is not accredited to ISO/IEC Guide 25 for calibration of the specific type of standard, it is the responsibility of the user of those standards to verify that the supplier has the management and technical procedures in place to carry out the calibrations competently.

Figure 2-4. How Does One Treat a Standard Lamp? Handling, Operation, and Storage of Calibrated Lamps. Calibrated lamps, as with all measurement standards, should be handled, operated, and stored with special care. Figure 2-4 summarizes key considerations in this regard.6 Additional information for specific measurement applications is available in the IESNA Lighting Measurements Testing and Calculation Guides (LM-5 through to LM-61). Other Radiometric Standards. Spectral standards are calibrated in terms of radiometric units. These are usually tungsten-halogen lamps, most often calibrated for spectral irradiance. The calibration typically includes portions of the UV and IR regions of the spectrum. Deuterium standards provide greater flux than tungsten-halogen lamps in the UV and extend the calibration range to below 250 nm.

General Methods

Photometric measurements, in general, make use of the basic laws of photometry previously described. Three types of photometric measurement procedures are: Direct Photometry. Direct photometry consists of the simultaneous comparison of a standard lamp and an unknown light source. Substitution Photometry. Substitution photometry consists in the sequential evaluation of the desired photometric characteristics of a standard lamp and an unknown light source in terms of an arbitrary reference. Relative Photometry. To avoid the use of standard lamps, the relative method is widely applied. It consists of the evaluation of the photometric characteristic of a lamp by comparison with the assumed lumen or spectral output of a test lamp.

MEASURING EQUIPMENT Radiometric measurement instrumentation consists of a detector, a means of conditioning or amplifying the output of the detector, a method of displaying or storing the measurement, and possibly an optical element or system of elements to collect the radiant quantity to be measured. Depending on the geometric relationship between the source and detector, the quantity measured is radiance, irradiance, or radiant intensity (or the corresponding photometric quantities: luminance, illuminance, or luminous intensity). A radiometer measures radiant power over a wide range of wavelengths that can include the UV, visible, and IR regions of the spectrum. It can employ a detector that is nonselective in wavelength response or one that gives adequate response in a specific wavelength band. Optical filtering can be used to level (flatten) the radiometer's response over a particular range of wavelengths or to approximate a desired function. For example, a number of useful action functions have been defined in the UV range, to which detector responses can be matched. Examples are industrial polymerization functions used for photoresist exposures and for UV curing. Another example is the erythemal effectiveness function (action spectrum) (see Chapter 5, Nonvisual Effects of Optical Radiation). The filtering must compensate not only for the spectral selectivity of the detector but also for the transmission characteristics of any optical components incorporated into the radiometer. Filtering also can be used to suppress the detector's response to radiant power outside the desired range. A radiometer that has been optically or electronically filtered to approximate a spectral sensitivity function of the fovea is called a photometer. The spectral response characteristic of a photometer is typically designed to match the CIE photopic standard observer. Photometers can also be filtered to provide a response similar to the CIE scotopic standard observer. A more elaborate radiometer is the colorimeter, which incorporates multiple detectors corrected to respond according to the CIE tristimulus functions. See Chapter 4, Color, for information on the CIE tristimulus functions. Filter colorimeters are used extensively to measure the color characteristics of visible radiation.

Detectors There is a broad range of detector options available. Choosing the correct detector will depend on the application in terms of spectral response, geometry, and quality. The characteristics of the signal, such as signal-to-noise ratio, amplitude, time response, and frequency bandwidth, all influence the choice of detector. The detector system's linearity range, field of view, noise equivalent power, and window transmission, as well as other factors, affect the measurement. Thermal Detectors. Thermal detectors include thermopiles, bolometers, and pyroelectric detectors. They produce a voltage proportional to the absorbed radiant power. The absorbing surface of the detector is usually blackened, making it nonselective over a wide range of wavelengths. The signal levels of these detectors are very low, and the detectors are very sensitive to ambient temperature changes. Once used extensively, they are now largely confined to laser light measurements. Phototubes. A phototube is a vacuum- or gas-filled glass tube containing a photoemissive surface as the source of electrical current. Photons striking the photoemissive surface release electrons by the photoelectric effect, and those electrons are collected by an anode having a higher voltage. The smallness of the resulting current limits the usefulness of vacuum phototubes to applications with high levels of radiant power. Adding a gas to a phototube and impressing a high voltage on the anode produces an avalanche amplification of the

current. Gas-filled phototubes can measure lower levels of radiant power; however, their nonlinear response to power makes them a poor choice for measurement purposes. The most useful form of phototube is the photomultiplier tube (PMT). PMTs employ a photocathode, which emits electrons when irradiated. The spectral sensitivity of a photomultiplier tube depends on the entrance window and photocathode material, for which many choices are available. Generally the photocathode has either a side-on or a head-on (or end-on) configuration. The side-on type is commonly used in spectroscopy applications and general photometric systems. The head-on type is commonly used for high-energy physics and scintillation counting. When photons strike the photocathode, electrons are emitted and then accelerated through a series of electron multipliers (dynodes), where the signal is greatly multiplied. The electrons are collected by an anode, where the output current is measured. A voltage divider chain connects the elements in the PMT in such a way that electrons are accelerated from one stage to the next. Typical PMT designs employ several to 15 stages of dynodes and produce signal gains from several thousand to hundreds of millions. The voltage required to operate the PMT can vary from 500 to 2000 V, depending on the tube construction and number of dynodes. The overall gain of the PMT is controlled by the voltage applied between elements. A high degree of voltage regulation is required for accurate operation. PMT detectors differ from solid-state devices in that they produce an output signal (dark current) in the absence of light, due to thermionic emission. The dark current can be reduced by lowering the temperature of the PMT. Most PMTs exhibit gain differences when exposed to magnetic fields or when their orientation in the earth's magnetic field is changed. Magnetic shielding is required in most applications. Most PMTs are shock sensitive, and rough handling can cause failure or loss of previous calibration. All phototubes have highly selective spectral response characteristics. Depending on the photoemissive cathode material used, a phototube can be used for UV, visible, or near-IR measurement; however, a single phototube cannot cover this entire spectral range.

Figure 2-5. The relative spectral sensitivity of selenium and silicon photovoltaic cells, the spectral distribution of tungsten radiation at 2856 K, and the spectral sensitivities assigned to the photopic standard observer. From A. Stimson, Photometry and radiometry for engineers. Copyright © 1974. Reprinted by permission of John Wiley & Sons, Inc. Solid-State Detectors. Solid-state detectors comprise a very large category of detectors incorporating semiconducting materials. All exhibit similar spectral response characteristics; their sensitivity to longer wavelengths increases up to a photon energy limit, where the detector response drops to zero. The useful spectral ranges of solidstate detectors extend from the UV to the far IR region. Some detectors are used in the photovoltaic mode, where the short-circuit current is measured; others are used as photoconductors, that is, a reverse bias voltage is applied and the device is treated as a radiation-sensitive variable resistor. Examples of photoconductive detectors include cadmium sulfide and cadmium selenide cells. The photovoltaic mode is most frequently used for radiant power measurements because of its inherent linearity of current as a function of incident radiant power level. The quantum nature of photovoltaic detectors makes them ideal for instruments that must perform over a wide range of radiation levels. The selenium barrier-layer cell, an early type of photovoltaic solid-state detector, was widely used in laboratory photometers but is now not recommended in photometers due to nonlinearity, fatigue, and instability. Today silicon photodiodes are commonly used in laboratory and commercial photometers. They offer a broader spectral range than the older selenium-based detectors (Figure 2-

5) and the ability to measure lower levels of radiant power. The silicon photodiode can be combined with a glass filter to match its spectral responsivity to the CIE photopic luminous efficiency curve. Silicon detectors are also used in self-scanning linear arrays, facsimile (fax) machines, spectral measuring instruments, and two-dimensional chargecoupled devices (CCDs). For more information see "Imaging Photometers" below. Photodiodes perform best when operated as current sources into "zero-impedance" amplifier circuitry. The linearity of silicon photodiodes has been shown to extend over 10 decades with appropriate amplification. Because very small currents are involved (typically 10−13 to 10−3 A), proper amplifier design is essential for the performance of these photometric instruments. Test methods, classes, and performance characteristics are discussed in more depth in Reference 18.

General Considerations Spectral Response. The detector is the primary component affecting the spectral response of a radiant-powermeasuring instrument. Photomultiplier tubes (PMTs) and silicon photodiodes are the most commonly used detectors in radiometers and photometers. As previously noted, these detectors respond differently to different regions of the spectrum. The spectral range of the detector should be matched to the spectral region to be measured. This can significantly improve sensitivity or relieve the burden of filtering. Photometers require good suppression of both UV and IR and good correction to the CIE luminous efficiency function. The CIE f1′ parameter is the only internationally agreed, illuminant-independent designation for the spectral error of photometers. Contrary to other designations in use, this one does not allow positive departures from V(λ) to cancel negative ones and is dependent of the spectral distribution of illuminant A. Transient Effects. Selenium photovoltaic cells, when suddenly exposed to constant illumination, require a short rise time to reach a stable output and thereafter can decrease slightly over a longer time due to fatigue. By contrast, silicon photodiodes typically exhibit microsecond rise times and no fatigue. The rise and fall times for most photometers employing silicon photodiodes are usually limited by the amplification circuitry. PMTs have nanosecond rise times but exhibit hysteresis, requiring from seconds to minutes to adapt to light-level changes. Precision radiometers and photometers usually employ PMTs with minimum hysteresis. Temperature Effects. Temperature variations affect the performance of all photodetectors. Selenium photovoltaic cells exhibit significant changes in shunt resistance with temperature, which can interact with external circuit impedance and produce gain changes. In addition, selenium cells can be permanently damaged by temperatures above 50°C (120°F). Silicon photodiodes are considerably less affected by temperature; however, problems can arise from the effects of temperature on detector response. The transmission of the spectral correction filters can also be affected by temperature. Correction factors can be employed when using photovoltaic detectors at temperatures other than their calibrated temperature (typically 25°C), or means can be provided to maintain the instrument temperature near the calibration temperature. Hermetically sealed detectors provide protection against the effects of humidity and some insulation against temperature cycling. Care should be taken that the effects of high temperature or temperature cycling do not damage cemented layers of the detector filter. PMTs are quite temperature sensitive. Both dark current and noise increase at higher temperatures. Also, the spectral response can vary significantly with temperature changes. Thermoelectric temperature control is frequently used to control the dark current, noise, and spectral characteristics of PMTs. Effect of Pulsed or Cyclical Variation of Light.7-13 Electric discharge sources flicker when operated on alternating-current (ac) power supplies (see "Flicker and Stroboscopic Effect" below). Precautions should be taken with regard to the effects of frequency, pulse rate, and pulse width when measuring the luminous properties of lamps.4,14-17 It cannot be assumed that an instrument will treat modulation of a light source in the same way as the human eye. The internal capacitance of the detector, the response time of the amplifier, and the response of the readout device (whether analog or digital) to pulsating signals must be considered. Special metering circuitry for the integration of pulsed light is available for the measurement of flashing incandescent and pulsed xenon sources. Instrument Zeroing. It is important to check the photometer or radiometer zeroing prior to taking measurements. For any type of equipment using an amplifier, it might be necessary to zero both the amplifier and the dark current. Where possible, it should be verified that the instrument remains correctly zeroed when the range is changed. Alternatively, any deviation from zero under dark current conditions should be measured and subtracted from the measurements. Electrical Interference. With electronic instrumentation, electrical interference can be induced in the leads between the detector and the instrumentation. This effect can be minimized by using filter networks, shielding, grounding, or

combinations of the above. Magnetic Fields. As previously noted, radiometers and photometers containing PMTs can be affected by strong magnetic fields. Commercial instruments containing PMTs use magnetic shielding adequate to protect them from most ambient magnetic fields; however, it is advisable to keep them away from heavy-duty electrical machinery. Signal Conditioning. The current produced by photovoltaic detectors usually requires amplification and other kinds of signal conditioning. The most common signal conditioning method uses an operational amplifier in a current follower configuration. This configuration provides low input impedance, and the output voltage is the product of the detector current and the feedback resistance. Although PMTs provide much higher signal currents, they are frequently used with a similar circuit to assure linearity. The output of the operational amplifier can drive most displays, whether analog or digital. Frequently, commercial instrumentation provides other kinds of signal conditioning. If signal currents are very low, integration can be used to increase the signal level and to improve the signal-to-noise ratio. Analog-to-digital conversion can be introduced in order to provide an interface to digital computing or a means of signal averaging. Memory may also be provided for data logging. Computing can be done inside the instrument by a microprocessor or by an external computer by means of a data link.

Measuring Instruments Instruments for photometric measurement are defined by their application. A photometric instrument can be used as a stand-alone system such as an illuminance or luminance meter, or combined with auxiliary equipment such as an integrating sphere to form a lamp measurement photometer system. CIE Publication 6918 serves as a guide for characterizing the performance parameters of photometers for luminance and illuminance (Figure 2-6a and 2-6b). Digital instrumentation for display and computer data acquisition further enhance the utility of modern photometric systems.

Figure 2-6, a and b, is exerpted from CIE Publication No. 69 (1987) and contains a summary of the expected errors of an illuminance meter and a luminance meter, respectively. Representative error values for best available commercial instruments have been estimated for several parameters. For all photometers the V(λ) match is of special importance, and should be as small as possible. Illuminance Meters Typical Configurations. The simplest illuminance meters consist of a photodiode with a photopic correction filter. The photodiode is connected to an operational amplifier with a display. These can be bench top, rack mountable, or portable. They can be enclosed in one case, or, as is more common with laboratory photometers, the detector and filter can be in one module that is connected by a cable to a console, at a convenient distance, containing the

amplifier and display. The electrical scheme can be anything from a simple amplifier with manual controls to a programmed microprocessor with routines for calibration, measurement, and conversion of display units. Some meters include communication ports for remote operation and data manipulation. Various commercial instruments are shown in Figure 2-7. Effect of Angle of Incidence (Cosine Effect). Illuminance meters are frequently used to measure the luminous flux density incident on a surface such as a table top, a wall, or a road surface. Part of the light reaching the detector at high angles of incidence is reflected by the cell surface or the filter or cover glass in front of it, and some may be obstructed by the rim of the case surrounding the detector. The resultant error increases with angle of incidence; where an appreciable portion of the flux comes at large angles, values as much as 25% below the true illuminance value can be obtained.

Figure 2-7. Various commercial illuminance meters. (a) Portable illuminance meter with small integrating sphere; (b) laboratory grade system that includes a photometer, a radiometer, and a fiber-optic power meter; (c) multifunction portable illuminance meter with detachable receptor head. The component of illuminance contributed by single sources at large angles of incidence can be determined by orienting the plane of the detector perpendicular to the direction of the light, and multiplying the reading thus obtained by the cosine of the angle of incidence. Detectors used in most illuminance photometers now have diffusing covers or some means of correcting the readings to a true cosine response. Solutions to the cosine problem include placing over the detector a flashed opal glass, diffusing acrylic disk, or an integrating sphere with a knife edge entrance port. With flashed opal glass and the diffusing acrylic disk at high angles of incidence, however, light will reflect specularly, so that the readings remain too low. This can be compensated by allowing light to enter through the edges of the diffuser. The readings at very high angles will then be too high but can be corrected by using a screening ring. The addition of auxiliary optics to improve cosine response can affect the photometric and directional response. CIE Publication 69 suggests correction methods for these errors.18 Leveling. Particularly during photometry of lighting systems where light is received from one or a small number of discrete sources, such as in roadway lighting, accurate leveling of the illuminance meter head is important. Instruments are available in which the detector is gimbal mounted and self-leveling. This removes problems when trying to measure horizontal illuminance on uneven or sloping surfaces.

Figure 2-8. Luminance meter with data processor.

Luminance Meters (Telephotometers) Luminance meters are essentially illuminance meters with the addition of suitable optics to image an object onto the detector (Figure 2-8). A means of viewing the object is usually provided so that the user can see the area that is being measured as well as the surrounding field. Because of the similarity of this optical system to a telescope, these instruments are also called telephotometers. Changing the focal length of the objective lens changes the field of view and thus the size of the measurement field. Some systems have apertures of various sizes to further define the measured area. Angular measurement fields from seconds of arc to several degrees can be selected. Typically, modern luminance meters use silicon photodiodes or PMTs. The amplifier sensitivity may be either manually selected or automatic. Color filters can be incorporated for color measurements, and neutral density filters to extend the dynamic range. Photodetectors are typically silicon for portable and low-sensitivity instruments and PMT for high-sensitivity instruments. Most instruments have at least a sensitivity dynamic range of four, and many incorporate attenuation screens or neutral-density filters for additional range. Most instruments in current manufacture incorporate digital displays. The electrical scheme can be anything from a simple amplifier with manual controls to a programmed microprocessor with routines for calibration, measurement, and conversion of display units. Some meters include communication ports for remote operation and data manipulation. Beamsplitter Spot Meters. This type of photometer employs, behind the objective lens, a beamsplitter, which divides the incoming radiation into two paths. Approximately half of the radiation passes through the beamsplitter and is focused on an aperture defining the measurement field. The radiation passing through the aperture can be measured with either a PMT or a solid-state detector. The radiation reflected from the beamsplitter is focused on a reticle having an etched pattern with the same dimensions as the measurement aperture. A viewing system with an eyepiece allows the user to see the field of view and an outline of the area being measured. The reticle must be carefully aligned with the measuring field. Readings are usually in cd/m2 or cd/ft2. Some instruments may include colorimetric filter options. Field-of-view capabilities may range from 0.25° to 10°, with sensitivity ranging from 10−2 to 106 cd/m2. Although good measurements can be made with this type of instrument, it does have some noteworthy disadvantages. Among these are loss of illumination to both the detector and the viewer; introduction of polarization, which affects the measurement of polarized sources (see Chapter 1, Light and Optics); and the difficulty of changing apertures and reticles for different measurement fields. In general, a low-cost instrument using a beamsplitter will provide adequate but not exact location of the measured spot. Aperture Mirror Photometers. Most of the problems of the beamsplitter spot meter are addressed by the aperture mirror photometer. There is no beamsplitter to introduce polarization error or reduce the brightness at either the measuring aperture or the viewed image. The image formed by the objective lens falls on an angled first surface mirror with a through hole for the measuring aperture. The viewing optics are focused on the aperture, which appears as a black circle. The field around the measurement aperture is clearly seen in the eyepiece. This arrangement allows apertures to be changed without the need to change precisely aligned reticles as well. A disadvantage of the aperture mirror photometer is that if a small source is imaged within the measuring aperture, it cannot be seen in the viewing optics. Instruments of this class usually employ high-quality detectors, one or more neutral-density range-multiplying filters, lens options, and some degree of colorimetric capability. They are available with internal microprocessor control and direct reading capability for luminance in several units, for color chromaticity coordinates, and for color temperature. The full-scale sensitivity for the best laboratory instruments ranges from 10−4 to 108 cd/m2. Imaging Photometers. Recent developments in imaging devices have provided a powerful tool for luminance measurements of complete scenes. Cameras equipped with a charge-coupled device (CCD) array are able to capture and digitize electronic images of visual scenes.19,20 Providing the proper controls are applied, the digital image can be used to determine the luminance at every point in the scene, corresponding to the pixels of the camera's CCD array. See Figure 2-9. A complete photometric capture can be carried out and saved in seconds. As the information is provided in digital form, complicated functions of luminance images can be analyzed and reported quickly for uniformity, contrast, spatial characteristics, and other photometric values. Some systems also provide chromaticity values. This form of photometry requires many factors to be controlled in the instrument and software if accurate results are

to be obtained. The CCD and optical attachments must be of high quality. "Field flattening" adjustments are required for all lenses that spatially distort the image to at least some degree. To measure the complete dynamic range of luminances in most interior scenes, earlier 8-bit systems required capture of multiple images at different exposure settings. Today, 16-bit systems are available that cover a dynamic range of over 65,000 : 1. For low luminance capability, cooling of the CCD array is required to reduce noise, to increase the detection limit, and to minimize susceptibility to changes in ambient temperature.

Figure 2-9. The battery-powered 496×288 pixel camera of a commercially available imaging photometric system with zoom lens and dynamic range of 0.05 to 500,000 cd/m2. The camera connects to a laptop computer by a parallel printer cable. The system can measure the luminances in an entire scene and automatically analyze, archive, and report the results. Applications for imaging photometers include the energy distribution of lamps (e.g., floodlamps and automobile headlights), production line quality control, luminance uniformity of a projected scene, and complicated analyses of scene illumination.

Reflectometers Reflectometers are reflectance measurement photometers. Reflectance measurements typically fall under three categories: diffuse, specular, and a mix of specular and diffuse reflectances. The design of the reflectometer and the method of measurement depend on the reflectance properties of the sample material and what part of the reflectance one desires to measure. ASTM 21 provides over seventy standards on color and appearance measurements, some of which employ reflectometers. Reflectance, the ratio of reflected light to incident light, is not simply a property of a material. Rather, it also depends on the measurement geometry, that is, the spatial relationship between the source and the detector. The fraction of the incident light reflected is very difficult to determine directly, particularly for diffuse reflection. To bring some order into what could be a chaotic measurement situation, reflectance is usually expressed as a reflectance factor, the ratio of the reflectance of a sample to that of a reflectance standard under the same measurement geometry. Three commonly used reflectance standards are a polished front surface mirror, a polished black glass having a specified index of refraction, and a total diffuse reflector (e.g., BaSO4). In one method commonly used for measuring total reflectance, the sample is illuminated by a narrow cone of light from a given angle, typically 10° or less from the normal to the sample surface, and the reflected light is collected over the entire hemisphere surrounding the sample. Instruments of this type are said to employ a conicalhemispherical geometry. The hemispherical flux collection is often accomplished by means of an integrating sphere with a detector, arranged so that it does not receive light reflected directly from the sample, but rather views the sphere wall. In this way the signal is proportional to the total flux reflected from the sample. The same type of instrument also can be used to measure only that part of the light that is diffusely reflected. One example of a sample that one might measure in this way is one with a very smooth dielectric surface that reflects strongly by scattering from pigments or other inclusions beneath the surface. In this case, light specularly reflected from the sample is allowed to escape through an specular subtraction port in the sphere wall, where a light trap can be positioned to absorb the specular reflected beam.

For measuring color, a 45/0 reflectometer is often used to evaluate the spectral character of diffusely reflected light. Figure 2-10a illustrates a typical mechanical layout. The source and the detector are mounted in a fixed relationship in the same housing. Light is incident on the surface from an angle of 45°, and the detector is positioned above and normal to the sample surface.

Figure 2-10a. A 45/0 reflectometry geometry. Reflectances from plane samples can be measured in many ways. One method employs a reflectometer that compares, with the aid of an auxiliary mirror, the incident flux with the flux after two reflections from the sample. Such a reflectometer is often available as an accessory to commercial spectrophotometers. Another method employs a goniophotometer that allows the user to position the light source and detector at any known angle. In some models, the sample holder can also be repositioned. Applications of the goniophotometer include measurements of gloss, luster, and haze. Another type22,23 of instrument, a Taylor Baumgartner sphere reflectometer, shown in Figure 2-10b, measures total reflectance. It consists of an integrating sphere, light source, and a photodiode. The sample is placed at the sample port of the integrating sphere. A collimated beam of light is directed onto the sample from approximately 30° to the normal, and the total reflected light, integrated by the sphere, is measured by the photodiode mounted in the sphere wall. The collimated light source is then rotated so that the light is incident on the sphere wall, and a second reading is taken. The sample is in place during both measurements, so that the effect on both readings of the small area of the sphere surface it occupies is the same.24 The ratio of the first reading to the second is the reflectance of the sample for the conditions of the test. Samples of translucent materials should be backed by a light trap. Various other instruments are available for measuring reflectance characteristics of materials.25-29 For any reflectance measurements, the reflectometer geometry employed should be specified, and for reflectance factor measurements the ideal reflector should also be specified.

Spectral Measuring Systems30 General Principles. The spectral response of a particular detector can be modified using optical or electronic filters to approximate some desired spectral response function (such as photopic correction). The detector itself, of course, must have adequate sensitivity over the spectral range being measured. Measurements with instruments that use corrected detectors are often called broadband or heterochromic radiometric measurements. Good-quality detectors are stable over time, and once they are calibrated, accurate measurements can be made without frequent corrections. Broadband instruments are very practical because they are inexpensive and simple to use. A disadvantage to broadband measurement is that it is difficult to design a filter correction to fit a desired function exactly, and although corrections can be applied, these corrections are usually themselves approximations. This can

be a problem for the most critical measurements that demand the highest accuracy. In some circumstances, measurement errors can be very large if the corrections are not appropriately applied or if there is a wide departure in filter correction from the ideal response function at wavelengths where a test source produces significant energy.3 Sometimes, as in the case of a photopically corrected detector, a very accurate fit can be achieved only at substantial cost, and even if the fit is initially satisfactory, the response of the detector or the filter can, at least theoretically, change over time. Commercial instruments usually provide an approximate correction, stating in their specification how closely their detector conforms to an ideal function. The only meaningful specification, however, is f1' (Figures 2-6a and b and CIE18), the only internationally accepted specification that does not allow negative departures from V (λ) to cancel positive departures.

Figure 2-10b. A Taylor Baumgartner sphere reflectometer. From R.S. Hunter, and R.W. Harold, The measurement of appearance, 2nd ed. Copyright © 1987. Reprinted by permission of John Wiley & Sons, Inc. Methods have been developed for correcting measurement errors due to imperfect filter design. For measurements on light sources, errors can be minimized by calibrating a detector with sources that have a known output and a spectral power distribution similar to that of the test source. There are also methods for characterizing a given detector and providing an analytical correction to the measurements.18 Perhaps the most important disadvantage of broadband measurement is the loss of specific wavelength information, resulting from the integration of radiation by the detector. Although detailed spectral information is not always needed, for some purposes the complete spectral power distribution (SPD) of a source, that is, the radiant power per unit wavelength as a function of wavelength, must be known. Spectral measurement systems are capable of determining the SPD in a very small band of wavelengths. The measurement of an SPD is considered fundamental; from these data, absolute radiometric, photometric, and colorimetric properties of a source can be determined. In comparison with broadband measuring systems, spectral measuring systems are complex and costly. The measurements also are generally more difficult and time consuming to make, often requiring a trained operator. There are different types of spectral measuring systems to suit specific applications, but they all generally incorporate the following elements: collection optics to receive and limit the radiation to be measured, a monochromator, a detector or detector array, electronics to process the detector or array signal, and some kind of readout or display. The monochromator houses a dispersing element such as a prism or diffraction grating that separates the various wavelengths of the spectrum. The monochromator has an entrance aperture, usually in the form of a rectangular slit, through which the collected radiation enters; maybe some optical elements that image the entrance slit onto the dispersing element(s); and an exit slit through which selected wavelengths of the dispersed radiation pass. A suitable detector or array is positioned at the exit slit to measure the SPD at the source. To make the many measurements necessary for a complete SPD, an automated system is recommended. There are many variations, on the automation

scheme, but typical scanning spectrometers incorporate a drive system to scan through a range of wavelengths (such as a means of rotating the monochromator grating) and a means of reading and storing the detector output for each sample, performing the necessary calculations and reporting the results (in printouts or graphs). This entire process is usually carried out using a computer. Detector array spectrometers acquire spectral data simultaneously without mechanical moving parts. Only a computer is needed for data processing. Types of Systems. A spectroradiometer is used to measure the SPD of light sources and relative spectral responsivities of detectors. Two methods are usually employed. For the first method, the collection fore-optics of a spectroradiometer system directs the radiation into a small integrating sphere or diffuser plate positioned in front of the entrance slit of the monochromator. This geometry is typically used to measure the spectral irradiance of a light source. For the second method, fore-optics of a spectroradiometer system directs radiance from a uniform source integrating sphere, an irradiated highly reflecting diffuse target, or a diffusely emitting lamp into the entrance slit of the monochromator. For the case of calibrating a detector, an appropriate lamp is chosen for its known spectral distribution. A spectrophotometer is used to determine spectral reflectance and transmittance properties of materials. Measurement results provide a means of examining the color of a material for analysis, standardization, and specification. In addition, it is the only means of color standardization that is independent of material color standards (always of questionable permanence) and independent of the differences in color vision existing among even socalled normal observers. Although called a spectrophotometer because of its principal application to measurements in the visible spectrum, this type of instrument is often designed for measuring UV and near-IR radiation. Some spectrophotometers are, in fact, designed specifically for UV or IR measurements. Spectroradiometers and spectrophotometers are closely related instruments in that they involve similar dispersion methods, detectors, and automation requirements. The principal difference is that a spectroradiometer performs measurements with the source(s) external to the system itself, whereas a spectrophotometer incorporates internal sources and an integrating sphere or chamber in which test samples are placed. It should be kept in mind that in using spectrophotometers for color standardization, it is really the reflectance properties of the test sample that are being measured and that the actual color appearance is dependent on the SPD of the source being used to evaluate the sample. Both of these instruments can be used in a spectrograph configuration. In this type of instrument, the exit slit is replaced by photographic film. Because there is no restrictive aperture, the dispersed radiation falls on the film plane and consequently the various wavelengths are spread out simultaneously. The exposed film provides a qualitative "picture" of the spectral components present. This is particularly useful in studying line emission sources, where the lines provide a "signature" of the source or material present. A means of measuring the optical density of the photographic emulsion where the various lines appear can provide quantitative information as to the intensity of the lines. This type of system obviates for mechanical scanning through the spectrum. In the form of a spectroradiometer, the instrument is used as an astronomical tool to evaluate the chemical composition of stellar objects. In modern instruments, the spectrograph employs diode arrays in the place of a photographic emulsion. Each diode detects incident radiation in a narrow wavelength band. Many modern spectroradiometers and spectrophotometers use this approach. Electronically scanned silicon photodiode arrays provide nearly instantaneous determination of a spectral power distribution. For many applications, array radiometry has replaced scanning systems, with the advantage of much greater measurement speed and the elimination of complex moving parts. For routine work, this reduced measurement time allows many more measurements to be taken, and changes in the SPD of a nonstable test source over time can be monitored. The disadvantages of array systems are that they inherently have more stray light, which is usually the limiting factor, and they do not have the absolute accuracy and sensitivity of the best scanning systems. Another very simple spectral measuring instrument, used to examine a spectrum visually rather than with a photodetector, is a spectroscope. Special Considerations. The ranges of spectral response in spectral measurement systems generally depend on the nature of the detector. Scanning spectroradiometers usually employ PMTs because of their high sensitivity. The response of PMTs extends from 125 to 1100 nm.31 Various types of silicon photodiodes cover the range from 200 to 1200 nm.31 For IR measurements several compounds can be used: intrinsic germanium (900 to 1500 nm), lead sulfide (1000 to 4000 nm), indium arsenide (1000 to 3600 nm), indium antinomide (2000 to 5400 nm), various types of doped germanium (zinc-doped, 2000 to 40,000 nm), and mercury cadmium telluride (1000 to 13,000 nm).32-33 The response of nonselective detectors spans a range from the near UV to beyond 30,000 nm.32,33 Where monochromators utilize diffraction gratings, the grating itself also influences the system response, so gratings must

be carefully selected for the range of wavelengths being measured. For accurate quantitative measurements, the electrical output of the detector must be known as a function of the input radiation to a spectral measuring system. Thus, for processing the electrical output of detectors (voltage, current, or charge), the instrumentation and the measurement method must be carefully selected with regard to a number of parameters, such as signal level (saturation), signal-to-noise ratios, fatigue, linearity, and response times (for rapidly varying signals). Photon counting and charge integration techniques are sometimes used for extremely low radiation levels.34,35 The influence of stray radiation on the measurement results must be minimized. This can be accomplished by designing an optical system to prevent unwanted wavelengths from reaching the detector. In all radiometric work there are two types of unwanted radiation: out-of-band radiation (radiation that is not completely dispersed) and higher-order radiation coming from diffraction gratings. Out-of-band radiation can usually be minimized by using a double monochromator (dispersing the radiation twice) or a single monochromator with appropriate filters. This type of unwanted radiation limits the use of diode-array spectroradiometers. It arises primarily from the properties of diffraction, so it cannot be eliminated entirely. Higher-order radiation can also be effectively limited by appropriate filters or by using a double monochromator employing a prism as one of the dispersing elements. Radiated flux of some wavelengths (mainly UV below 200 nm) is dispersed or absorbed by a layer of air between the radiator and the detector. In this case consideration must be given to the placement of the source and the detector, and to the surrounding medium. All observed spectroradiometric data (sometimes called raw data) are a function not only of the SPD of the light source but also of the spectral throughput of the optical system, the spectral bandwidth of the monochromator, and the spectral responsivity of the detector. Collectively, these define the system responsivity, which is determined by measuring at each wavelength the output of a calibration or standard source having a known output. Once this function is known, the test-source observed data can be corrected by multiplying each value by the known output of the standard source divided by its observed value at that wavelength.

Photometric Measuring Systems: Basic Equipment Types Optical Bench Photometers. Optical bench photometers are used for the calibration of instruments for illuminance measurement. They provide a means for mounting sources and detectors in proper alignment and a means for easily determining these relative distances between them. If the source is of known luminous intensity in a specified direction and is distant enough from the detector so that its radiation can be treated spatially as if it were emanating from a point, the inverse square law (Equation 2-1) can be used to compute illuminance. Distribution Photometers. For characterizing the spatial distribution of illumination from a source, a series of luminous intensity measurements are made on a distribution photometer, which can be one of the following types:

Goniometer and single detector Fixed multiple detector Moving detector Moving mirror

All types of distribution photometers have advantages and disadvantages. The significance attached to each advantage or disadvantage is dependent on other factors, such as available space and facilities, polarization requirements, and economic considerations. Goniometer and Single Detector. The light source is mounted on a goniometer, which allows it to be rotated about both horizontal and vertical axes. The luminous intensity is measured by a single fixed detector. There are several different versions of goniometers. Each is related to the type of source or luminaire being measured and the facilities in which it is located. With the use of computers, the coordinate system of a goniometer system can be easily transformed to another coordinate system;36 thus consistent data-reporting formats become practical. Figures 2-11a, 2-11b, and 2-11c show three types of goniometer systems, known as Type A, B, and C. Details are provided in LM-35-1989.37 Types B and C are most commonly used for outdoor and indoor sources, respectively. Note that these designations differ from the Types A, B, and C photometry defined by the CIE.

Figure 2-11a. Type A goniometer with fixed horizontal axis: (a) related coordinate systems; (b) representation on a sphere of the X-Y coordinate system; and (c) X-Y coordinate system.

Figure 2-11b. Type B goniometer with fixed vertical axis: (a) related coordinate system; (b) representation on a sphere of the V-H coordinate system; and (c) V-H coordinate system.

Figure 2-11c. Type C goniometer with a photocell or mirror movable around a horizontal axis. Fixed-Multiple-Detector Photometer. Numerous individual detectors are positioned at various angles around the light source under test. Readings are taken on each detector to determine the intensity distribution. See Figure 2-12. Moving-Detector Photometer. This device consists of a detector that rides on a rotating boom or arc-shaped track; the light source is centered in the arc traced by the detector. Readings are collected with the detector positioned at the desired angular settings. Sometimes a mirror is placed on a boom to extend the test distance. See Figure 2-13. Moving-Mirror Photometer. This is a Type C photometer in which the mirror rotates around the light source, reflecting the light to a single detector. Readings are taken at each desired angle as the mirror moves to that location. See Figure 2-14.

Figure 2-12. Schematic side elevation of a fixed multiple cell photometer. Integrating-Sphere Photometer. The integrating-sphere photometer is used to measure the total luminous flux from a source (lamp or luminaire). The most common type is the Ulbricht38 sphere. The theory of the integrating sphere assumes an empty sphere whose inner surface is perfectly diffusing and of uniform nonselective reflectance. Every point on the inner surface then reflects to every other point, and the illuminance at any point is therefore made up of two components: the flux coming directly from the source and that reflected from other parts of the sphere wall. With these assumptions, it follows that the illuminance, and hence the luminance, of any part of the wall due to reflected light only is proportional to the total flux from the source, regardless of its distribution. The luminance of a small area of the wall, or the luminance of the outer surface of a uniformly diffuse transmitting window in the wall, when carefully screened from direct light from the source but receiving light from other portions of the sphere, is therefore a relative measurement of the total luminous flux from the source. Figure 2-15 shows the Ulbricht-type integrating sphere with a high-reflectance, diffuse white interior. Diffuse coatings of lower reflectance are also used. There are advantages as well as disadvantages to the level of reflectance of the sphere coating or material used. An integrating sphere coated with an 80% diffuse reflectance coating is less susceptible to spectral errors due to surface contamination than that from a sphere coated with a diffuse high-reflectance coating. The compromise is reduced efficiency. If a reference source of known output (in terms of luminous flux) is measured in the integrating sphere, a calibration constant can be determined, and thus the luminous flux of a source of unknown output can be determined.

Figure 2-13. Schematic diagram of a moving cell photometer. The presence of a source having finite dimensions, its supports and electrical connections, the necessary baffles or shields, auxiliary accessories, and the exit window or ports are all departures from the basic assumptions of the integrating-sphere theory. While durable high-reflectance diffuse material (Figure 2-15) and coatings are now available for sphere interiors, none exhibits the ideal properties of perfect diffusivity and spectral nonselectivity. Despite these limitations, if the reference source and the test source are similar in shape, size, surface reflectance characteristics, and light distribution patterns, the errors introduced by a imperfect integration can be small. For accurate measurements of sources dissimilar from the reference source, corrections must be applied for self absorption, spectral mismatch, and spatial nonuniformity, which are inherent with integrating sphere lamp measurement photometry.4,39-43

Figure 2-14. Schematic diagram of a moving mirror photometer.

Figure 2-15. Ulbricht-type integrating sphere with diffuse high-reflectance coating, shown here at the National Institute of Standards and Technology (NIST). Alternatives to the Ulbricht-type integrating sphere exist.44 Recently, an alternative integrating-sphere method using an external source has been developed at the National Institute of Standards and Technology (NIST) (Figure 2-16).5 In this geometry the total luminous flux φv,i of a source in the integrating sphere is calibrated against an external reference source calibrated for illuminance, at an aperture outside the integrating sphere. The total luminous flux φv,e of the external source can be determined from Ea and aperture area A.

Figure 2-16. Alternative geometry for luminous flux measurements. Despite its shortcomings, the integrating sphere is an important tool for photometric measurements. It can produce results quickly with a single measurement, and suitable accuracy can be achieved if good instrumentation and proper procedures are employed.

LABORATORY MEASUREMENTS4,45-48 Precision and accuracy of laboratory measurements can be consistently attained by correctly following a set of standard procedures and using good equipment. Such procedures are prepared and published by standardizing committees or consensus organizations to serve as a common basis for measurements. IESNA guides provide

detailed procedures for the many required electrical and photometric tests for different lamp types. These should be consulted when testing or calibration is being performed. The following information is a general overview of good laboratory practices. Important elements that standard procedures embody are:

Controlling the electrical supply characteristics Ensuring electrical stability of reference and working standards Securely mounting the source Minimizing the influence of stray light Frequently checking instrument readings Stabilizing sources and auxiliaries by operating and maintaining them for a sufficient period of time before measurements are taken Compensating for the inherent nonsymmetry of sources and luminaires48 Using instruments that have adequate precision and accuracy to meet the requirements of the test Understanding the limitations of instruments used with respect to sensitivity, linearity, and dynamic range Ensuring that the instruments are suitably calibrated and maintained

The laboratory must provide an adequate test distance for sealed beam, floodlight, and projector lamp measurements. Space can be conserved by using mirrors to fold the effective test distance. Large integrating spheres can require extending the ceiling height. There must be adequate space surrounding the sphere to gain access to the sphere interior. Temperature and air circulation control are critical requirements for discharge lamp photometry. Cleanliness, provision of suitable electric power, and storage space for lamps, luminaires, and instruments are also important requirements in any photometric laboratory, as are the provision of suitable training and the maintenance of records of the response or output of instruments and reference standards. These and other general guidelines are described in ISO/IEC Guide 25, General Requirements for the Competence of Calibration and Testing Laboratories.49 See "Traceability and Accreditation," earlier in this chapter.

Electrical Measurements Photometric results depend on the electrical operating characteristics of the source being tested. For this reason, electrical and photometric measurements are almost always made concurrently, and both sets of data are included in the report. Electrical measuring instruments should be selected to have current and voltage ratings corresponding to the circuit conditions to be encountered and should give indications of the desired precision and accuracy. Digital instrumentation has replaced analog instruments in most cases, and their use is highly recommended. A comprehensive discussion on instrumentation requirements is given in the IES guide "Selection, Care, and Use of Electrical Instruments in the Photometric Laboratory."50 If analog instruments are used, it is especially important to consult this reference so that appropriate instrument corrections can be applied for the type of measurement being performed. Test lamps must be stable before attempting any accurate electrical measurement. It is necessary to season the test lamps in accordance with established procedures.51 Instrumentation Direct-Current Circuits CURRENT. Current-measuring instruments, or ammeters, are inserted in series with the source and have low impedance, thus adding little to the load on the power supply. They can be self-contained, but this type of instrument might not have sufficient range to directly measure the current through high-wattage sources. When current exceeds several amperes, it is usually determined using an accurate digital voltmeter (DVM) in combination with a calibrated current shunt. The shunt resistance should be sufficient to produce a voltage drop that can be measured accurately, although too high a resistance will lower the current appreciably. For most lamp types, an appropriate shunt resistance is between 0.1 and 1 Ω. VOLTAGE. The voltage applied to the light source is measured by means of a voltmeter connected in parallel with the source. It should have the highest impedance possible so as not to disturb the circuit. To avoid corrections to compensate for a voltage drop in the ammeter or across the current shunt, the voltmeter is usually connected directly across the load. Often, separate voltage leads are connected to the base of the lamp through special lampholders (kelvin sockets) to avoid voltage drop errors resulting from socket-to-lamp connections. POWER. Power can be computed as the product of current and voltage. Alternating-Current Circuits. Instruments that measure alternating current must be compatible with the waveform and frequency of the voltage or current being measured. They should have a frequency response comparable to the supply frequency. For measurements of sinusoidal waveforms, instruments that indicate rms values are satisfactory. For accurate measurements of distorted waveforms that contain harmonics, true rms instruments with frequency response capabilities well above the fundamental frequency are required. In some ac circuits, a dc component can also be present. In this case a true rms instrument that measures ac and dc must be used; otherwise a separate dc measurement is necessary to determine the true rms value of the voltage or current. Many measuring instruments do not include the dc component in the measurement, and this feature needs to be verified.

As described for dc instruments, ac instruments are connected to the test circuit, and the same impedance considerations apply. In electrical measurements with high-intensity discharge lamps, some instruments require protection from transient voltages (on the order of 1500 to 4500 V) that occur when lamps are turned on and off. This is usually accomplished by providing switches to connect the instruments into the circuit only after the lamp is operating. CURRENT. Ammeters for use in ac circuits can be self-contained; however, as with dc circuits, the rating of many meters is too low. A transformer or ac current shunt can be used in combination with a digital meter to achieve a more extensive measuring range. The shunt should be noninductive to avoid current waveform distortion and phase shifting. For this reason, dc current shunts are generally unsuitable. VOLTAGE. As with dc circuits, voltage can be measured directly across the load using an ac voltmeter. The measuring range can be extended using a potential transformer or voltage divider. POWER. In a resistive circuit (e.g., an incandescent filament lamp) the ac power load can be either computed as the volt-ampere product or measured directly with a wattmeter. In reactive circuits (e.g., a circuit with a magnetic ballast or where distorted waveforms are present), power is measured with a wattmeter, which measures or computes the real power by averaging all the instantaneous volt-ampere products over one cycle of lamp operation. Such an instrument should be capable of responding to harmonics well above that of the fundamental. Older analog wattmeters require corrections to achieve high accuracy. Modern digital power-measuring instruments are often capable of performing all the necessary electrical measurements and seldom require corrections. The specification of the instrument should be checked to ensure the instrument is capable of accurately measuring the specific waveforms.

Electrical Measurements of Incandescent Lamp Circuits Incandescent filament lamps usually are measured on a dc circuit where accurate measurement results can be obtained using inexpensive dc power sources and measuring instruments. High-accuracy electrical measurements of incandescent lamps are especially desirable because lamp output is very sensitive to small changes in electrical settings. For example, a 1% change in current through a typical filament lamp results in a 5 to 7% change in light output. Some incandescent sources have integral electronic components such as diodes and are designed to operate only on ac circuits. True rms instruments that also measure the dc component of a waveform should be used for measurements on this type of source. Electrical Measurements on Discharge Lamp Circuits. All electric discharge lamps have negative volt-ampere characteristics and must therefore be operated in conjunction with internal or external current-limiting devices, such as resistors or reactors. These are described in Chapter 6, Light Sources. Because of the presence of distorted waveforms, true rms measuring instruments must be used for any measurements on discharge lamp circuits. Such measurements may involve lamps or ballasts. In some cases the two are inseparable and measurements are made on the combination as a single device. Because of normal manufacturing tolerances, commercial ballasts supply lamps with some variation in voltage and current characteristics, which affect the electrical input and the light output of lamps. To promote uniformity of testing, the International Electrotechnical Commission (IEC), working through the American National Standards Institute (ANSI), the Canadian Standards Association (CSA), and similar national standardizing bodies throughout the world, has established or is establishing standardized testing procedures for determining the electrical characteristics for most of the common types of discharge lamps. These standard tests are performed using reference circuits and reference ballasts that comply with specified electrical requirements. Where international standards have not been established, national standards are used. Lamp Testing. Lamp parameters are influenced by many factors. Detailed, accepted testing procedures where these factors are controlled or specified are described in the appropriate IESNA guides.52-54 Some of the more important conditions affecting lamp test results are listed below:

Ambient temperature Drafts Lamp position Lamp connections Lamp stabilization Power-supply characteristics Ballast characteristics Lamp circuit characteristics

Ballast Testing. Ballast parameters are influenced by many factors. Detailed accepted testing procedures are described in the appropriate ANSI standards.55-60 Ballast testing requires consideration of some or all of the following. VOLTAGE RANGE. For most tests, ballasts should be operated at their rated primary voltage. REFERENCE LAMPS. Some tests on ballasts specify that the ballast shall be operating a reference lamp. Reference lamps are seasoned lamps that, when operated under stated conditions with the specified reference ballast, operate within specified tolerances of electrical values established by the appropriate existing or proposed specifications. OPEN-CIRCUIT VOLTAGE. This measurement is necessary only for ballasts containing a transformer. ELECTRODE HEATING VOLTAGE. On ballasts for use with lamps having continuously heated electrodes, the electrode heating voltages are measured with the electrode windings loaded with a specified dummy load. SHORT-CIRCUIT CURRENT (ballasts for high-intensity discharge lamps). An ammeter is inserted in the circuit in place of the lamp, and the short-circuit current of the ballast is measured. STARTING CURRENT. Ballasts for instant-start fluorescent lamps, a resistor and ammeter, in series, with a total resistance equivalent to the value specified in the appropriate standard,55-61 is used instead of the lamp. For ballasts used with high-intensity discharge lamps, the secondary circuit is short-circuited. ELECTRODE PREHEATING CURRENT (preheat ballasts for fluorescent lamps). This measurement is made with an ammeter connected in series with the lamp electrodes while the lamp is maintained in the preheat condition. BALLAST OUTPUT FOR FLUORESCENT LAMPS. For preheat and instant start ballasts, specifications are in terms of the power delivered to a reference lamp operated by the ballast under test, as compared with the power delivered to the same lamp by the appropriate reference ballast. With continuously heated electrodes, specifications are in terms of the light output of a reference lamp operated with the ballast under test, as compared with the light output of the same

reference lamp when operated with the appropriate reference ballast. BALLAST REGULATION FOR FLUORESCENT LAMPS. Relative lamp power input and light output are measured at 90% and 110% of rated ballast input voltage. FLUORESCENT LAMP CURRENT. The current of a reference lamp should be measured on both the ballast under test and the reference ballast. For lamps with continuously heated electrodes,61 unless the internal connections of the ballasts are accessible, measurement of lamp current requires special instrumentation to supply the vector summation of currents in the two leads to an electrode.

Photometric Measurements Incandescent Filament Lamps.62

In determining the photometric characteristics of bare incandescent filament lamps, the requirements for electrical measurements previously described should be observed. Test lamps (except series types) are usually measured at rated voltage. Reference lamps63-65 can be purchased from or recalibrated by NIST, NRC, or other established national or commercial laboratories. Reference lamps are usually rated for lumens at a current or voltage a little below their nominal rating in order to extend the burning time. The correct color temperature and lamp filament temperature should be maintained. Nickel-plated bases are used on these standards to reduce corrosion and high-resistance problems over their life. Working standards usually are calibrated against reference standards. They should have the loops of filament supports closed firmly around the filament to avoid the possibility of random short-circuiting of a portion of the filament by the support. They should be adequately seasoned and selected by successive comparisons with reference standards for stability. All standards should be handled carefully to avoid exposure to electrical and mechanical shocks. Exposure to current or voltage above the standard value may alter lamp ratings. For a more in-depth discussion on photometric standards see Reference 42. It is recommended that the voltage applied to the test lamp be ramped up slowly to its final setting (Figure 2-4). Intensity Measurements. Sources can be measured on an optical bench photometer if either the luminous intensity in a particular direction or a mean horizontal luminous intensity is desired. Lamps standardized for unidirectional measurements are usually marked to indicate the orientation. A common practice is to inscribe a circle and a vertical line on opposite sides of the bulb. The standardized direction is from the circle toward the line, when they are centered on each other, looking toward the receiver. Total Flux Measurements. Most routine photometric measurements on incandescent filament lamps are for total light output or total luminous flux and are made in a sphere (Figure 2-15). Best results are obtained when the standard lamp has approximately the same physical size, lumen output, color temperature, and location in the sphere as the test lamp. Lamp depreciation measurements usually are taken at 70% of rated lamp life. By this time, some blackening of the bulb is likely. This blackening can lead to errors in photometric measurements taken with an integrating sphere because the blackened area of the lamp absorbs some of the interreflected light. To overcome these errors, a third lamp, commonly called the "absorption," "comparison," or "auxiliary" lamp, should be installed in the sphere so that it is shielded from both the integrating sphere detector and the test lamps. Successive readings should be taken with the absorption lamp operating: first with the reference (known) lamp installed but not operating, then with the blackened (aged) lamp installed but not operating. The difference between these readings represents the light absorbed by the blackened lamp and can be used to correct the values given by the integrating sphere. The same general procedure can be followed in most cases where the characteristics of the integrator are altered during the test by the introduction of light-absorbing elements (see the above discussion on "Integrating-Sphere Photometer"). Photometry of Discharge Lamps.66-68 As with incandescent lamp measurements, the photometric characteristics of discharge lamps usually are determined in conjunction with electrical measurements, whose general requirements have been given. The substitution method is normally employed for photometric measurements. Complete, detailed photometry procedures can be found in IESNA test guides.52-54 Equipment BALLASTS. When a lamp is measured for rating purposes, it should be operated on the appropriate reference ballast. If no standard exists, the ballast should comply with the general lamp requirements. In general practice, photometric measurements of fluorescent lamps burning on commercial ballasts should be made with the ballast operating at rated input voltage, and measurements on high-intensity discharge lamps should be made with the lamp operating at rated wattage. The ballast should be operated long enough to reach thermal equilibrium. The use of commercial ballasts should conform to the procedures given in the appropriate standards.47 DETECTORS. Detectors should be selected according to the criteria given in "Illuminance Meters" above. Additional corrections may have to be applied to the measured data. STANDARD LAMPS. These lamps should have characteristics similar to the lamp under test with respect to light output, physical size, shape, and spectral distribution. INTEGRATING SPHERES.38,44 The integrating sphere to be used should comply with the requirements described in the previous discussion of integrating-sphere photometers. To provide acceptable performance, integrating spheres should be of adequate size for the lamp being tested. Direct substitution is not always possible, and generally the larger the integrator with respect to the test lamp dimensions, the smaller any necessary corrections will be. Also, the ambient temperature in a larger sphere is less affected by heat generated by the test lamp. The sphere diameter should be at least 1.5 m for high-intensity discharge lamps and at least 1.2 times the length of the lamp for straight lamps; the area of the light source should not exceed 2% of the interior surface of the sphere. If direct substitution is used, these requirements are less stringent. DISTRIBUTION PHOTOMETER.69-71 The lamp is mounted in open air with the distance between receiver and lamp at least five times the lamp length or 3 m, whichever is greater. Except as stipulated below, the lamp should be operated in the same burning position as associated with the luminaire for which it is intended, and it should be held stationary during measurement. Movement can disturb its stabilization.52-54 The total light output can be computed if the lumen-candlepower ratio is known, or if strict substitution is practiced. The measured luminous intensity values for a lamp are established by multiplying the test readings by the photometric calibration constant. The total light output is the sum of products of these values with the appropriate zonal lumen constants. Measures should be taken to exclude stray light, to control ambient temperature and drafts, and to reduce the effects of lightabsorbing or -reflecting materials. For fluorescent lamp measurements the lamp is mounted in a horizontal position and measurements taken normal to the axis of the lamp. To provide the greatest accuracy, these measurements must be taken at several angular positions by rotating the lamp around its axis between sets of measurements. For measurements of high-intensity discharge lamps, especially metal halide lamps, the lamp should be placed in its designed operating position. If this is vertical, measurements can be made while the lamp is rotating slowly on its longitudinal axis. Holding the lamp stationary is desirable for stabilization and accuracy. If the lamp is to be operated in any other position, it must be held stationary during measurement, since the light distribution from a high-intensity discharge lamp is a function of arc position, which is influenced by gravity. Reflector-Type Lamps.72 For purposes of identification, a reflector-type lamp is defined as a lamp having a reflective coating applied to part of the bulb, this reflector being specifically contoured for control of the luminous distribution. Included are pressed or blown lamps such as PAR and ER lamps, as well as other lamps with optically contoured reflectors. Excluded are lamps of standard bulb shape to which an integral reflector is added, such as silvered-bowl and silvered-neck lamps; lamps designated for special applications, such as automotive headlamps and picture projection lamps, for which special test procedures are already established; lamps having translucent coatings, such as partially phosphor-coated mercury lamps; and reflector fluorescent lamps. Intensity Distribution. Several methods for measuring intensity distribution are available, depending on the type of lamp and the purpose of the test. The photometric center of the lamp is usually taken as the center of the bulb face, disregarding any protuberances or recesses in the face. The test distance should be great enough so that the inverse square law applies (Equation 2-1). The intensity distribution of a circular beam is commonly represented by an average curve in a plane along the beam axis. (The beam axis is that axis around which the average distribution is substantially symmetrical; the beam axis and photometric axis are adjusted to coincide.) The curve is obtained either by taking measurements with the lamp rotating about the beam axis, or by averaging a number of curves (at least eight) taken in planes at equally spaced azimuthal intervals about the axis.

The intensity distribution of a lamp whose beam is oval or rectangular in cross section is not adequately represented by one average curve. For some lamps, two curves through the beam axis, one in the plane of each axis of symmetry, can supply sufficient information. The necessary number of traverses, their distribution within the beam, and the intervals between individual readings vary considerably with the type of lamp; sufficient measurements should be made to describe the average distribution pattern adequately. When reflector-type lamps are considered for a specific application, test results will be most readily comparable when in the same form as that for equipment used for the same application. For example, when a direct performance comparison of a reflector lamp with floodlighting luminaires is desired, the lamp should be tested according to approved floodlight testing procedures.37 The same is true for indoor luminaire applications. Total Luminous Flux Measurements.4 The total luminous flux can be obtained by direct measurement in an integrating sphere or by calculation from intensity distribution data. Because of the high-intensity spot produced by most reflector-type lamps, special precautions should be taken when using an integrating sphere. One possible position for the test lamp in the sphere is with its base close to the sphere wall and the beam aimed through the sphere center, thus distributing the flux over as large an area of the sphere as possible. An appropriate baffle should be placed between the light source and the detector. When reflector-type standards are available, the calibration of the sphere follows the usual substitution procedure, and for maximum accuracy the standard lamp should be of the same type as the test lamps. Beam and Field Flux. The beam and field flux can be calculated from an average intensity distribution curve or from an isocandela diagram. Of particular interest is the flux contained within the limits of 50% and 10%, respectively, of the maximum intensity. The beam angle is defined as the total angular spread of the cone intercepting the 50%-of-maximum intensity. The field angle is defined as the total angular spread of the cone intercepting the 10%-of-maximum intensity. Flicker and Stroboscopic Effect.73 All light sources operated on alternating current will flicker. The degree to which flicker is perceived, if at all, depends on the frequency of the alternating current, the persistence of light generated by the source, and the viewing conditions. Flicker has special significance for objects moving within the field of view. Objects may appear to move discretely rather than continuously under flickering illumination; this is known as the stroboscopic effect. The magnitude of the effect depends on the rate and amplitude of the flicker, the rate of object motion, and the viewing conditions. The flicker index73 has been established as a reliable relative measure of the cyclic variation in output of various sources at a given power frequency. It takes into account the waveform of the light output as well as its amplitude. It is calculated by dividing the area above the line of average light output by the total area under the light output curve for a single curve (Figure 2-17). Area 2 in Figure 2-17 may be close to zero if light output varies as periodic spikes. The flicker index assumes values from 0 to 1.0, with 0 for steady light output. Higher values indicate in increased possibility of noticeable stroboscopic effect, as well as lamp flicker.

Figure 2-17. Curve of the light output variation from a lamp during each cycle, showing the method of calculating the flicker index.

Spectroradiometric Measurements of Light Sources30,74 Spectroradiometric measurements of light sources can be reported in relative terms or as absolute values. The units of the latter are generally watts per unit area as a function of specified wavelength bands for spectral irradiance, and watts within the wavelength bands for spectral flux. Extended light sources typically are measured for their spectral radiance in units of watts per unit area per solid angle per unit wavelength. Spectroradiometric measurements provide the fundamental data for the determination of radiant quantities; all other quantities can be computed from these measurements. Measurement Methods. Spectroradiometric measurement involves the comparison of two light sources: a reference source of known SPD and a test source whose SPD is to be determined. The two sources can be compared wavelength by wavelength with a scanning monochrometer, or they can be measured sequentially by completely scanning each source throughout the spectrum. When the sources are compared wavelength by wavelength, the two sources are operated simultaneously. This approach necessitates discontinuous wavelength scanning, pausing at each measurement point. It also requires that the reference and test source not drift over the lengthy measurement period. The advantage of this method is the minimizing of errors due to shortterm drift in the instrument's response. When the sources are measured sequentially, either continuous or discontinuous wavelength scanning can be adopted. When this method is used, the spectral response of the measuring system must remain constant throughout the measurement period. Instrument drift can be checked by remeasuring the standard source. With discontinuous measurements, several readings at each wavelength can be taken and averaged, or other filtering methods can be employed to minimize the effect of electrical noise on the measurement. With continuous measurements, the spectrum is scanned from one end to the other at a uniform rate, and the signal from the detector is integrated over discrete intervals within the scanning period. This approach effectively converts the spectral power distribution into a histogram, the height of a segment being proportional to the power emitted by the source over that interval. In principle, this method does not require that the scanning intervals and the monochromator bandwidths be perfectly correlated, but the best resolution is obtained when they

are. An advantage of this method is excellent rejection of electrical noise, which is integrated out of the measurement in the signal averaging that takes place over each interval. It is also efficient, as there is no stopping and starting of the wavelength drive. When this method is adopted, the timing requirements are critical and the system response time must be adequate to capture the rapidly changing signal levels encountered during the scanning process. This is especially important when strong spectral lines are present. Modern spectroradiometers incorporate a means of automating both the operation of the system and the data collection process.75,76 Some form of immediate presentation of the data, such as a graph on a chart recorder, is desirable to aid in the recognition of malfunctions in the measurement system or instability in the test source. The final presentation of the data is usually in the form of a series of values that describe the SPD of the light source over a particular spectral range. Typically, an SPD curve accompanies the report of numerical data on output versus wavelength. A suitable wavelength interval for a particular test is chosen by considering the type of spectra being measured. Excessively small intervals are wasteful, requiring the collection of unnecessary data. Excessively large intervals can result in missed information. For optimum results, the sample interval and the bandwidth of the instrument are the same. The SPD from an incandescent lamp can be adequately represented by a series of measurements at 10-nm intervals across the spectrum, since the intensity changes gradually with wavelength. For measurements on discharge lamps, whose spectra include a number of emission lines, a smaller interval is required. For the calculation of chromaticity, which is one of the most important applications of spectroradiometric data, a 2-nm sample interval and bandwidth gives values of the chromaticity coordinates x and y accurate to ±0.001 for almost any type of source, and this is adequate for most purposes.

Color Appearance of Light Sources For measurement of the color appearance of a light source, see Chapter 4, Color.

Life Performance Testing of Lamps Life tests are performed on a very small portion of the products under consideration. Under such conditions, test program planning, sampling techniques, and data evaluation become especially important.77-79

It is not practical to test lamps under all of the many variables that occur in service; hence specific reproducible procedures must be included in the test experiment plan. Incandescent Lamp Life Testing.80-83 Life tests of incandescent filament lamps can be divided into two classes: rated-voltage and overvoltage tests. Rated Voltage. Lamps are operated in the specified burning position at a voltage or current held within ±0.25% of rated value. Sockets should be designed to assure good contact with lamp bases, and the racks should not be subjected to excessive shocks or vibration. If lamps are removed for interim photometric readings, care should be taken to avoid accidental filament breakage. Sockets should be lubricated, because vibration can break a filament that has been rendered brittle by burning. Overvoltage (Accelerated) Tests. Lamp life is shortened by voltages in excess of rated. Extreme overvoltage life testing, sometimes called "high forced testing," exponentially shortens lamp life so that lamps can be tested in less time. The exponents are empirical and require many comparison tests at rated voltages to determine them. Electric Discharge Lamp Life Testing.84-86 Tests generally are made using ac supply. The power supply should have a voltage waveform in which the harmonic content does not exceed 3% of the fundamental. The line voltage should be regulated. There is no widely accepted method of accelerated life testing of discharge lamps. Auxiliaries. Since an electric discharge lamp must be operated with auxiliaries, which often affect lamp life, they must be selected to conform to the requirements of the appropriate guides, test methods, and specifications. Test Cycles. An on-off cycle is normally employed to simulate field conditions. The commonly accepted cycles are 3 h on and 20 min off for fluorescent lamps, and 11 h on and 1 h off for high-intensity discharge lamps, although others are in use. It is known that more rapid cycling (that is, shorter off-times) shortens lamp life, but the correlation with the standard cycle is not sufficiently accurate to predict the life on the standard cycle. Environment. Vibration, shock, room temperature, and drafts should be controlled to minimize their effect on measurement results. Orientation. Lamps should be tested in an orientation recommended by the lamp manufacturer. See the IESNA guides for testing specific lamp types.37,52-54,62

Luminaire Photometry The purpose of making photometric measurements of a luminaire is to determine its light distribution and characteristics in a way that will most adequately describe its performance. Characteristics such as intensity distribution, zonal lumens, efficiency, luminances, beam widths, and typing are necessary in designing, specifying, and selecting lighting equipment. Photometric data are essential in deriving and developing additional application information. The information that follows is only a rudimentary guide to the photometry of luminaires. Specific photometric guides and practices are referenced below and should be consulted to obtain the detailed testing procedure for each type of luminaire. The IES Practical Guide to Photometry47 provides information covering general photometric practices, equipment, and related matters. Each specific type of luminaire (indoor lighting, task lighting, floodlighting, or streetlighting) requires different testing procedures. However, there are several general requirements that should be met in all tests. The luminaire to be tested should be (1) typical of the unit it is to represent, (2) clean and free of defects (unless it is the purpose of the test to determine the effects of such conditions), (3) equipped with lamps of the size and type recommended for use in service, and (4) installed with the light source in the recommended operating position for service. If the location of the source in a beam-producing luminaire is adjustable, it should be positioned as recommended to obtain such a beam as is desired in service. To provide an accurate description of the characteristics of the materials used in the manufacture of a luminaire, measurements should be made of the reflectances of reflecting surfaces where applicable. Luminaires should be tested in a controlled environment under controlled conditions. The photometric laboratory temperature should be held steady. Typically, for fluorescent photometry, where lamps are sensitive to temperature variations, the room temperature should be held to 25 ± 1°C. Power supplies should be regulated and free of distortion to minimize any effects of line voltage variations. Test rooms should be painted black or provided with sufficient baffling to minimize or eliminate extraneous and reflected light during testing. For accurate measurements, the distance between the luminaire and the light sensor should be great enough that the inverse square law applies (Equation 2-1). The minimum test distance is governed by the dimensions of the luminaire. This distance should not be less than 3 m (10 ft) and at least five times the maximum dimension of the luminous area of the luminaire. For best accuracy the test distance should be measured from the center of the apparent source to the surface of the detector. However, from a practical standpoint, the following rules should suffice: (1) for recessed, coffered, and totally direct luminaires, the test distance should be measured to the plane of the light opening (plane of the ceiling); (2) for luminous-sided luminaires the test distance should be measured to the geometric center of the lamps; (3) for suspended luminaires, (a) if the light center of the lamp(s) is within the bounds of the reflector and there is no refractor, the test distance should be measured to the plane of the light opening, (b) if the light center of the lamp(s) does not fall within the bounds of the reflector and there is no refractor, the test distance should be measured to the light center of the lamp(s), and (c) if a refractor is attached, then the test distance should be measured to the geometric center of the refractor.

General-Lighting Luminaires Intensity Distribution. See the following IESNA guides for specific information on testing general-lighting luminaires: Photometric Testing of Indoor Fluorescent Luminaires,87 Photometric Testing of Indoor Luminaires Using High Intensity Discharge or Incandescent Filament Lamps,88 and Reporting General Lighting Equipment Engineering Data.89,90 The basic measurement made in a photometric test of a luminaire is the luminous intensity in specified planes and angles. The resulting intensity distribution is used to determine zonal lumens, efficiency, and average luminances. It is therefore essential that sufficient data be taken to adequately describe the intensity distribution and the luminaire's total luminous flux. Luminaires having a symmetric distribution can be measured in five to twelve equally spaced planes and the results averaged. Most fluorescent luminaires are measured in five planes per quadrant in opposite quadrants, and the results for the two quadrants averaged to give the five-plane data. To adequately describe a highly asymmetric luminaire it might be necessary to measure in planes at 10° (or smaller) intervals. The distribution in each vertical plane is determined by taking readings at 10° (or smaller) intervals. If the luminaire is of the beam-forming type, the intensity measurements should be made at smaller intervals within the beam-forming area. If visual comfort probability calculations are to be made, it is recommended that intensity measurements be made at least at every 5° in the vertical plane, and preferably every 2 ½°. Most luminaires are measured using the relative method, similar to that for lamps. From these readings, total luminous flux can be estimated using the zonal lumen method. The lumen rating of the lamp is divided by this value to give the constant factor for that lamp on that photometer. By multiplying by that constant, the readings taken on the luminaire can be converted to luminous intensity for the lamps operating as rated. The intensity distribution data generally are presented in tabular form on the test report sheet. Data for a lens or indoor luminaire are given in five planes. Three distributions (parallel to the lamps, 45° to parallel, perpendicular to the lamps) usually are presented in the form of polar distribution curves. Luminance. Either before or after the photometric test, while the lamps are still installed and stabilized, the maximum luminance of the luminaire should be measured at the angles specified in the appropriate guide and at the shielding angles. The readings should be taken both crosswise and lengthwise in the case of fluorescent luminaires or luminaires giving asymmetric distributions. The projection of the measurement field should be circular and of 6.45 cm2 (1 in.2) area. The luminance photometer must be calibrated for the SPD of the test lamps. If average luminance values are desired, they can be calculated by using the intensity measurements obtained from the test data. By definition, luminance is the luminous intensity (candlepower) of any surface in a given direction per unit of projected area of the surface viewed from that direction.91 Total Luminous Flux. The total luminous flux from of the luminaire, needed to establish its efficiency in terms of the total luminous flux of the installed lamps(s), can be determined in an integrating-sphere photometer or by computations from the intensity distribution data. If it is to be measured in a sphere, the efficiency can be determined by the relative method (Figure 2-18). First, lamps are mounted at the center of the sphere and a reading taken. A reading is then taken on an auxiliary lamp mounted at another point in the sphere. The luminaire is then mounted at the center of the sphere, and a reading is taken. Another reading is then taken with the auxiliary lamp mounted in the sphere. The efficiency is then calculated as follows:

where R1 = reading of lamp(s) at the center of the sphere, R2 = auxiliary lamp reading, R3 = luminaire reading, R4 = auxiliary lamp reading with luminaire in the sphere. Only one lamp (or one luminaire) should be operated whenever a reading is being taken. The sphere method is not considered to be as accurate as the method using luminous intensity distribution data. Intensity distribution data are used to compute the luminous flux in any angular zone from nadir to 180°. The product of the midzone intensity and the zonal constant gives the zonal lumens. The summation of the zonal lumens multiplied by 100 and divided by the nominal lamp lumens gives the efficiency in percent. Constants useful in calculating luminous flux from intensity data are given in Figures 2-19 and 2-20. For computing the luminous flux, the average luminous intensity at the center of each zone should be multiplied by the zonal constant (Figure 2-19) equal to 2π (cos θ1 - cos θ2). In Figure 2-18, the zone limits represent θ1 and θ2. The measurements should be made at the midpoint of this interval. The constants in Figure 2-20 are computed for intensity measurements on projector-type luminaires made on a goniometer of the type shown in Figure 2-11b. In this figure, the vertical spacing is φ and the horizontal angle and setting represents the midpoint between θ1 and θ2. If the measurements have been made with the type shown in Figure 2-11a, the same constants can be used by interchanging the vertical and horizontal angular arguments, that is, by substituting the word "vertical" wherever "horizontal" appears, and vice versa. The zonal constants for Figure 2-20 were computed as φπ (sin θ2 - sin θ1)/180, where φ is the vertical interval and θ1 and θ2 are the limits of the horizontal interval. If a goniometer of the type in Figure 2-11a is used, the constant is equal to θπ (sin φ2 - sin φ1)/180, where θ is the horizontal interval and φ1 and φ2 are the limits of the vertical interval. See Reference 91 for additional zonal constants.

Figure 2-18. Luminaire and lamp positions within an integrating sphere for the relative method of determining luminaire efficiency. The diameter of the sphere should be at least twice the maximum dimension of the luminaire to be measured.

Figure 2-19. Constants for Use in the Zonal Method of Computing Luminous Flux from Intensity Data

Figure 2-20. Continued

Figure 2-20. Continued

Figure 2-20. Continued

Figure 2-20. Constants* (K) for Converting Beam Intensity of Projector-Type Luminaires (Searchlights, Floodlights, and Spotlights) into Luminous Flux If a number of constants are to be calculated for the same interval, the following shortcut method is useful and accurate. For the first formula above, let θm be the midzone angle and let P equal one-half the zone interval. The formula becomes

The zone width is often the same for a series of constants. In these cases, the first factor is simply multiplied successively by the sines of the midzone angles. For the second formula above, let θm be the median angle on the horizontal interval. The formula then becomes

It is commonly assumed that a ballast operated at its rated input voltage delivers rated wattage to a lamp, and that a lamp operated at its rated wattage delivers rated total luminous flux. In many instances, the assumption is invalid. Therefore, a procedure has been developed92 to provide a factor called the equipment operating factor to be applied to photometric data on luminaires to adjust them to the specific combination of luminaire, lamp type, and ballast used in a system. By repetitive tests, this procedure can be used to determine variations of system performance exclusive of lamp variations. Such a factor can be applied specifically to total luminous flux, intensity, and illuminance as given on photometric data sheets. Two possibilities are recognized: the first (LLB1) in which the lamp is used in the operating position for which it is rated, and the second (LLB2) in which the lamp is operated in a position other than the one for which it is rated. The factors determined are equipment-specific (luminaire-lamp-ballast combination) under initial conditions unless specified otherwise. The procedures essentially involve a relative measurement of total luminous flux. The test luminaire is operated at the rated supply voltage, after operating conditions have stabilized. Relative flux measurements are obtained with the test ballast and, without extinguishing the lamp, with the reference ballast operated at its rated input voltage. The equipment operating factor is the ratio of the first to the second measurement of total luminous flux.

Floodlight-Type Luminaires. The following applies to floodlighting equipment having total beam spread (divergence) of more than 10°. For specific information on testing this type of equipment, consult the IES Approved Method for Photometric Testing of Floodlights Using Incandescent Filament or Discharge Lamps.37 For equipment having a beam spread less than 10°, see the IES Guide for Photometric Testing of Searchlights.93 The classification of floodlights is based on horizontal and vertical beam width. The classification is designated by National Electrical Manufacturers Association (NEMA) type numbers.94 For symmetrical beams the floodlight type is defined by the average of the horizontal and vertical beam spreads. For asymmetrical beams it is defined by the horizontal and vertical beam spreads in that order; for example, a floodlight with a horizontal beam spread of 75° (Type 5) and a vertical beam spread of 35° (Type 3) would be designated as a Type 5 × 3 floodlight. Stray light is defined as light emitted by the floodlight that is outside the classified beam. In some instances, stray light is useful in illumination. It can also be detrimental, depending on its magnitude and direction. To determine the amount and direction of stray light, it is necessary to make measurements as far horizontally and vertically as the readings have significant values in relation to the measuring system. If the geometric center of the emitted light is not enclosed by the reflector, the floodlight should be mounted on the goniometer so that the light center of the lamp is at the goniometric center. If the lamp light center is within the reflector, the floodlight should be positioned so that the center of the reflector opening coincides with the goniometric center. Either the direct or the relative method of photometry can be used for floodlights. The relative method has an advantage in that cumulative errors can be reduced and maintenance of standards of luminous intensity and flux is not necessary. In the latter method, relative intensity readings for the test lamp alone made with a distribution photometer, and for the lampfloodlight combination made with a floodlight photometer, are taken with the lamp operating under identical electrical conditions in both tests. The method of taking intensity readings is to traverse the beam with such angular spacings as to give approximately 100 uniformly spaced readings throughout the beam. For the definition of the beam limit, see Reference 95. By interpolating between these readings, an isocandela diagram can be plotted on rectangular coordinates (see Chapter 9, Lighting Calculations). The total luminous flux in the beam can be computed using the constants in Figure 2-20. The information usually reported for floodlights includes the following: NEMA type, horizontal and vertical beam distribution curves, maximum beam luminous intensity, average maximum beam luminous intensity, beam spread in both horizontal and vertical directions, beam flux, beam efficiency, total floodlight flux, and total efficiency. The report should also indicate whether the data were obtained with a Type A, Type B, or Type C goniometer (Figure 2-11a, b, and c).96,97 Roadway Luminaires. A guide has been prepared to provide test procedures and methods of reporting data to promote consistent evaluation of roadway luminaires performance that use incandescent filament or high-intensity discharge lamps.98 Luminaires selected for test should be representative of the manufacturer's product. A test distance of 8 to 10 m (25 to 30 ft) should be sufficient for most beam-forming luminaires. The photometric test distance is generally defined as the distance from the light center to the surface of the detector, taking into account the distances to and from any mirrors that might be used. The number of planes explored during photometric measurements should be determined by the symmetry or irregularity of the distribution and by the purpose of the test. The number of vertical angles at which readings are taken depends on how the readings are to be used. If an isocandela diagram is to be plotted, readings might have to be taken at close intervals, especially if the values change rapidly. Computer acquisition systems provide comprehensive evaluation of luminaires and lighting application designs; readings taken at vertical angle intervals through the beam section do not exceed 2½°. For luminaires having a distribution that is symmetrical about a vertical axis (IES Type V), readings should be taken in ten or more vertical planes and averaged. For luminaires having a distribution that is symmetrical about a single vertical plane (IES Types II, III, IV, and II four-way), readings should be taken in vertical planes that are no more than 10° apart. To simplify data processing, it can be advantageous to divide the beam section laterally into 10° zones and measure at the midzone angle. Averages may be taken of the readings at corresponding angles on the opposite sides of the plane of symmetry. Any computations that are to be performed can then be done on one side of the plane of symmetry, using the averaged data. For luminaires having a distribution that is symmetric about two vertical planes (IES Type I), readings should be taken as above, but the computations may be performed in one quadrant of the sphere. For luminaires having a distribution that is symmetric about four vertical planes (IES Type I four-way), readings may be taken as above, but the computations may be performed in one octant of the sphere. For luminaires having an asymmetric distribution, readings should be taken in vertical planes that no more than are 10° apart. Since there is no symmetry, any computations performed should be done without averaging the data obtained from different planes. Sufficient data should be obtained to allow classification of the light distribution in accordance with recommended practice (see Chapter 22, Roadway Lighting) as well as to provide an isolux (isofootcandle) diagram, the utilization efficiency, and the total and four-quadrant efficiencies. Projector Luminaires.99,100 The equipment required for photometric measurements of projectors, searchlights, and beacons is similar in most respects to that required for other types of photometry. When the projector luminaire is of unusual size and weight, it might be necessary to use its own mounting and goniometric facilities for the photometric work, or to hold the luminaire fixed and traverse the beam by moving the detector. Since projector luminaires can have a total beam spread of less than 1° and furthermore can be massive, the mechanical requirements on the goniometer are severe (Figure 2-21). Rigidity, freedom from backlash, and accuracy of angular measurements are prime requirements. There should be provision for accurate angular settings of the order of 0.1° or less.

Figure 2-21. Typical goniometer for indoor photometric range. Two rotary tables of the type used in large machine tools are incorporated into the goniometer to provide two of the three rotations available. For horizontal distributions either the rotary table on which the outer frame is mounted or the inner table on which the test equipment is mounted may be used. This makes for flexibility and is especially useful for obtaining polar angle distributions. The inner frame with the test unit mounted can then be balanced with the adjustable counterweights. After balancing, a small constant torque is applied to the inner frame through the use of the pulley and weight at the side, thus eliminating backlash. Indoor photometric measurements are preferred but frequently impracticable because of the lengths required. For photometry on the relatively short indoor ranges, proper photometric procedures must be followed.31,43,92,101 Outdoor ranges require much more attention to methods of reducing stray light, minimizing atmospheric disturbances, and correcting for atmospheric transmission. Range sites should be selected where the terrain is flat and uniform. The range should be as high off the ground as practicable. Stray light should be minimized by a suitable system of diaphragms. Any remaining stray light should be measured and subtracted from all readings.

Figure 2-22. Diagram showing distances and dimensions used to determine minimum inverse-square distance. Ranges should not be located where atmospheric disturbances occur regularly or where dust or moisture are prevalent. The absorption and scattering of light by moisture, smoke, and dust particles, even in an apparently clear atmosphere, can introduce considerable errors in measurements.102,103 Therefore, it is desirable to measure the atmospheric transmittance before and after the test. A standard reference projector is frequently employed, calibrated either by repeated observations in the clearest weather, when the atmospheric transmittance can be accurately estimated or independently measured, or by laboratory measurement methods. The illuminance from searchlights, beacons, or other highly collimating luminaires, if measured at distances greater than a certain minimum, obeys the inverse square law (Equation 2-1). The minimum distance is a function of the focal length of the reflector, the diameter of the reflector aperture, and the diameter of the smallest element of the light source (arc stream or filament). This minimum distance is called the beam crossover point and is the distance where the optic is seen to be completely flashed, that is, the minimum distance where the refracting lens or the reflector is seen as completely luminous. Only at distances greater than this does the inverse square law apply. This minimum distance can be calculated by using the following general formula (Figure 2-22):

where L0 = minimum distance for optic under consideration, a = distance from the optical axis to the outermost flashed point when viewed from a distance point on the optical axis, d = distance from the centroid of the light source to the same outermost flashed point as used to determine a, s = diameter of the smallest element of the light source (for example, the arc stream width of an arc source, or one coil of a multicoil filament lamp), K = constant equal to 500 when a, d, and s are in mm and L0 is in m; and 6 when a, d, and s are in inches and L0 is in ft. The above calculation determines the minimum inverse square distance based on ideal light sources and axial measurements, and therefore should be considered approximate. In practice, the range used should be much larger than this calculated distance to ensure conformance. Methods for using shorter ranges and "zero-length photometry" have been devised; however, the full length range gives the highest accuracy. These methods and a fuller discussion of minimum inverse-square-distance calculations can be found in References 92, 100, and 104.

Luminance Measurements105

Luminance measurements of lamps and luminaires should be made by either the absolute or the relative method. With the absolute method, reference standards must be available for equipment calibration. In practice the relative method is generally used. The published luminances of symmetric fluorescent lamps are computed from the rated lumen output of the lamp according to the following formula:106

where Lavg is the average luminance of the full width of the lamp at its center, in candelas per square meter. The diameter and length are expressed in meters or inches, and K for T-12 lamps is as follows:

To compute the approximate luminance Lθ

where Kθ

of a fluorescent lamp at any angle to the lamp axis, the following formula is used:

is as shown in Figure 2-23.

Figure 2-23. Values at Various Angles of the Lamp: Intensity Ratio Kθ

for Preheat-Starting Types of Fluorescent Lamps (Average for 15, 20, 30, 40, and 100 W Lamps)

A luminance meter should be used to make direct measurements of the luminaire luminance. The characteristics of the luminance meter should be such that the field of measurement corresponds to a projected area of 645 mm2 (1 in.2) normal to its axis at the distance of measurement.107 This can be achieved with an appropriate lens system and a measurement distance determined from the manufacturer's specifications. This technique has the advantage of the observer being able to view the exact area being measured through the luminance photometer.

Calorimetry of Luminaires A thermal testing method has been developed for compiling data on air-cooled heat transfer luminaires.108

The method uses a calorimeter to measure the thermal energy distribution of the luminaires. The entire laboratory room in which the calorimeter is located becomes a part of the calorimetric system. The room must be controlled closely with respect to temperature, air motion, and relative humidity. Varying conditions can affect the results of the calorimetric measurements. The room conditions should be as follows: the temperature should be controlled at 25 ± 0.3°C; the velocity of the air in the space containing the calorimeter should be held constant and not exceed 0.15 m/s (30 ft/min); the relative humidity should be held constant at any convenient value between 20 and 50%; and the room should not be affected by external conditions. The selection of a calorimeter type is determined by the purpose of the device and the degree to which its conditions can be controlled. Three types of calorimeters are the zero-heat-loss calorimeter, a calorimeter constructed to compensate for the heat transfer through its walls; the calibrated-heat-loss calorimeter (the approved IESNA type), a box in which the heat loss can be determined by dissipating a measured quantity of energy in the plenum; and the continuous-fluid-flow calorimeter, a modification of the zero-heat-loss calorimeter consisting of a heavily insulated heat exchanger installed over the luminaire. Precision instrumentation is needed to measure temperature (thermometers, thermocouples, thermistors, and resistance elements), pressure (manometers, micromanometers, draft gauges, and swinging-vane gauges), mass flow rate of air and water, electrical quantities, and luminous flux. Each luminaire that is to be tested for energy distribution should first be measured photometrically in accordance with accepted procedures. During calorimetry, the photometer should be installed at luminaire nadir, not less than 300 mm (1 ft) from the bottom of the enclosure. The distance at which the detector should be mounted below the luminaire is limited by the distance required for the cell to integrate flux over the entire luminous area of the luminaire. It is necessary that precautions be taken to prevent the detector from responding to luminous flux other than that transmitted by the luminaire under test. Its position must not be changed during the test. The data to be recorded and reported should include the description and size of the luminaire, mode of operation, test conditions (space and plenum temperatures), relative light output from a 25°C base with the luminaire operated in free air outside the calorimeter, energy to space, energy removed by the exhaust air stream, and exhaust air temperature, all as functions of the exhaust flow rate. Energy and relative luminous flux should be reported as a function of the air volume flow rate.

FIELD MEASUREMENTS In evaluating an actual lighting installation in the field it is necessary to measure or survey the quality and quantity of lighting in the particular environment.

Field measurements apply only to the conditions that exist during the survey. Recognizing this, it is very important to record a complete detailed description of the surveyed area and all factors that might affect results, such as interior surface reflectances, lamp type and age, voltage, and instruments used in the survey. In measuring illuminance, detectors should be cosine and color corrected ( Figure 2-6).109 They should be used at a temperature above 15°C (60°F) and below 50°C (120°F), if possible. Care should be exercised while taking readings to avoid casting shadows on the detector of the measuring instrument, and also by standing far enough away from the detector, especially when wearing light-colored clothes, to prevent light from the source from being reflected onto it. A high-intensity discharge or fluorescent system must be lighted for at least 1 h before making measurements. In new lamp installations, at least 100 h of operation of a gaseous source should elapse before measurements are taken. With incandescent lamps, seasoning is accomplished in a shorter time (20 h or less for common sizes). The IESNA has developed a uniform survey method for measuring and reporting the necessary data for interior applications.110 The results of the uniform surveys can be used alone or with those of other surveys for comparison purposes, to determine compliance with specifications, and to reveal the need for maintenance, modification, or replacement. The IESNA survey method was compared with other survey methods and with comprehensive illuminance measurements on horizontal working planes in 11 different rooms under actual conditions of use.110 The rooms varied in room cavity ratio and in luminaire configurations. They included five of the six room types shown in Figure 2-24. The study found that the IESNA method was generally reliable to within an accuracy of 10%. Larger errors can be expected for spaces with unusual room cavity ratios or poor uniformity. It was concluded that, of all the survey methods studied, the IESNA method has the advantage of requiring the smallest number of measurement points and has no problems of coincidence of luminaire and measurement grids because the measurement points are fixed to the luminaires. The rigidity of the method is a disadvantage, however, in spaces that are obstructed, lack orthogonal geometry, or have highly nonuniform illumination. In these cases, a measurement grid can be used.

Interior Measurements Illuminance Measurements--Average Determination of Average Illuminance on a Horizontal Plane From General Lighting Only. The measuring instrument should be positioned so that when readings are taken, the surface of the detector is in a horizontal plane and 760 mm (30 in.) above the floor. This can be facilitated by means of a small portable stand to support the detector. Daylight can be excluded during illuminance measurements by taking the readings either at night or with shades, blinds, or other opaque covering on the fenestration. The area should be divided into approximately sized squares, taking a reading in each square and calculating the arithmetic mean. A measurement grid of 0.6 m (2 ft) is suitable for many spaces. For spaces with unusual room cavity ratios or highly nonuniform illumination, as in corridors under emergency lighting conditions, see Chapter 9, Lighting Calculations. Ouellette et al.111 have shown that nonuniform lighting, as, for example, in some emergency lighting applications, may require several hundred randomly selected measurement points to obtain a representative value for average illuminance. This study illustrates how difficult it is to obtain an accurate representative value for illuminance in highly nonuniform illuminated spaces. Regular Area with Symmetrically Spaced Luminaires in Two or More Rows. See Figure 2-24a. 1. Take readings at stations r-1, r-2, r-3, and r-4 for a typical inner bay. Repeat at stations r-5, r-6, r-7, and r-8 for a typical centrally located bay. Average the eight readings. This is R in Equation 2-8. 2. Take readings at stations q-1, q-2, q-3, and q-4 in two typical half bays on each side of the room. Average the four readings. This is Q in Equation 2-8. 3. Take readings at stations t-1, t-2, t-3, and t-4 in two typical half bays at each end of the room. Average the four readings. This is T in Equation 2-8. 4. Take readings at stations p-1 and p-2 in two typical corner quarter bays. Average the two readings. This is P in Equation 2-8. 5. Determine the average illuminance in the area by using the equation

where N = number of luminaires per row, M = number of rows. Regular Area with Symmetrically Located Single Luminaire. See Figure 2-24b. Take readings at stations p-1, p-2, p-3, and p-4 in all four quarter bays. Average the four readings. This is P, the average illuminance in the area. Regular Area with Single Row of Individual Luminaires. See Figure 2-24c. 1. Take readings at stations q-1 through q-8 in four typical half bays located two on each side of the area. Average the eight readings. This is Q in Equation 2-9. 2. Take readings at stations p-1 and p-2 for two typical corner quarter bays. Average the two readings. This is P in Equation 2-9. 3. Determine the average illuminance in the area by using the equation

where N = number of luminaires.

Regular Area with Two or More Continuous Rows of Luminaires. See Figure 2-24d. 1. Take readings at stations r-1 through r-4 located near the center of the area. Average the four readings. This is R in Equation 2-10. 2. Take readings at stations q-1 and q-2 located at each midside of the room and midway between the outside row of luminaires and the wall. Average the two readings. This is Q in Equation 2-10. 3. Take readings at stations t-1 through t-4 at each end of the room. Average the four readings. This is T in Equation 2-10. 4. Take readings at stations p-1 and p-2 in two typical corners. Average the two readings. This is P. 5. Determine the average illuminance in the area by using the equation

where N = number of luminaires per row, M = number of rows. Regular Area with Single Row of Continuous Luminaires. See Figure 2-24e. 1. Take readings at stations q-1 through q-6. Average the six readings. This is Q in Equation 2-11. 2. Take readings at stations p-1 and p-2 in typical corners. Average the two readings. This is P in Equation 2-11. 3. Determine the average illuminance in the area by using the equation

where N = number of luminaires. Regular Area with Luminous or Louvered Ceiling. See Figure 2-24f. 1. Take readings at stations r-1 through r-4 located at random in the central portion of the area. Average the four readings. This is R in Equation 2-12. 2. Take readings at stations q-1 and q-2 located 0.6 m (2 ft) from the long walls, at random lengthwise of the room. Average the two readings. This is Q in Equation 212. 3. Take readings at stations t-1 and t-2 located 0.6 m (2 ft) from the short walls, at random crosswise of the room. Average the two readings. This is T in Equation 212. 4. Take readings at stations p-1 and p-2 located at diagonally opposite corners 0.6 m (2 ft) from each wall. Average the two readings. This is P in Equation 2-12. 5. Determine the average illuminance in the area by using the equation

where W = number of luminaires per row, L = number of rows.

Figure 2-24. Location of illuminance measurement stations in (a) regular area with symmetrically located single luminaire; (b) regular area with symmetrically located single luminaire; (c) regular area with single row of continuous luminaires; (d) regular area with two or more continuous rows of luminaires; (e) regular area with single row of continuous luminaires; (f) regular area with luminous ceiling.

Figure 2-25. Form for Tabulation of Point Illuminance Measurements

Illuminance Measurements--Point With task, general, and supplementary lighting in use, the illuminance at the point of work should be measured with the worker in his or her normal working position. The measuring instrument should be located so that when readings are taken, the surface of the light-sensitive cell is in the plane of the visual task or of that portion of the visual task on which the critical visual processing is required--horizontal, vertical, or inclined. Readings should be recorded as shown in Figure 2-25.

Figure 2-26. Form for Tabulation of Luminance Measurements

Luminance Measurements Luminance surveys should be made under actual working conditions and from a specified work point location with the combinations of daylight and electric lighting facilities available. Consideration should be given to sun position and weather conditions, both of which can have a marked effect on the luminance distribution. All lighting in the area--task, general, and supplementary--should be in normal use. Work areas used only in the daytime should be surveyed in the daytime; work areas used both day and night should be surveyed under both conditions, as the luminance distribution and the possibilities of comfort and discomfort can differ markedly between them. Nighttime surveys should be made at night or with shades drawn. Daytime surveys should be made with shades adjusted to positions actually set by the occupants. On a floor plan sketch of the area, an indication should be made of which exterior wall or walls, if any, were exposed to direct sunlight during the time of the survey by writing the word "Sun" in the appropriate location. Readings should be taken, successively, from the worker's position at each work point location A, B, C, etc., and luminance readings from each location recorded as shown in Figure 2-26.

Outdoor Measurements In roadway and many floodlight installations, light is projected to the surface to be lighted, and each luminaire must be adjusted carefully to produce the best utilization and quality of illumination. For an accurate evaluation of this type of installation, illuminance must be measured carefully. Preparation for the Survey112-114

1. Inspect and record the condition of the luminaires (globes, reflectors, refractors, lamp positioning, etc.). In the case of roadway lighting, make sure the luminaires are level and their vertical and lateral placement is as designed. Unless the purpose of the test is to check depreciation or actual in-service performance, all units should be cleaned and new lamps installed. New lamps should be seasoned properly.95-97 While inoperative lamps are readily noticed in roadway installations, they can easily

be overlooked in large floodlighting systems. If these lamps are not replaced for the field survey, proper consideration must be given when evaluating the results. 2. Measure and record the mounting heights of the luminaires. 3. Measure and record the locations of the poles, the number of luminaires per pole, the wattage of the lamps, and other pertinent data. Check these data against the recommended layout; a small change in the location or adjustment of the luminaires can make a considerable difference in illuminance. 4. Determine and record the burning hours of the installed lamps. 5. Consider the impact of stray light on the measurements. The survey should be made only when the atmosphere is reasonably clear. Extraneous light produced by a store, a service station, or other lights in the vicinity requires careful attention in street lighting tests. 6. Luminaire voltage should be measured because it can affect illuminance. At night, during the hours when the luminaires are normally used, record the voltage at the lamp socket with all of the lamps operating. The voltage at the main switch can be used instead, provided allowance is made for the voltage drop to the individual luminaires. If discharge lamps are used, record the input voltage to the ballast at the ballast terminals. Discharge lamps should be operated at least 60 min to reach normal operating conditions before measurements are made. Survey Procedures. Measurements should be made with a recently calibrated, color-corrected, and cosine-corrected photometer (Figure 2-6) capable of being leveled for horizontal measurements or positioned accurately for other measurement planes as required. The photometer should be selected for its portability, repeatability, and measurement range. 1. For roadway lighting systems, at least one traffic lane must be closed for substantial periods of time. Because of this difficulty and expense of making field measurements of pavement luminance, it is common to use a computerized design procedure using point calculations of horizontal illuminance level at each of the pavement luminance measurement points recommended. As a check on the computer calculations, it is necessary only to measure the illuminance at a reduced number of points.112 2. For roadway signs, the minimum and maximum illuminance levels are determined by scanning the sign face. Additional illuminance measurements are taken at specific locations according to the sign size. Luminance measurements are also made for both externally and internally illuminated signs.114 3. For sports installations,113 the sports area (or the portion of the area under immediate consideration) should be divided into test areas of approximately 5% of the total area, and readings should be taken at the center of each area. Where lighting for color television is involved,115 horizontal illuminance readings need to be taken at each station. Another set of readings should be made with the detector tilted 15° from vertical in the direction of each camera location and 0.9 m (36 in) above ground level (unless otherwise specified for the particular activity). This will help ensure adequate illumination for the color television cameras. 4. Readings should be made at each test station, with repeat measurements at the first station frequently enough to assure stability of the system and repeatability of results. Readings should be reproducible within 5%. Enough readings should be taken so that additional readings in similar locations will not change the average results significantly.

REFERENCES 1. Commission Internationale de l'Éclairage. 1983. The basis of physical photometry, CIE publication no. 18.2. Paris: Bureau Central de la CIE. 2. National Bureau of Standards. 1991. The International System of Units (SI). 6th edition. NBS Special Publication 330. Gaithersburg, MD: National Bureau of Standards. 3. Ouellette, M.J. 1993. Measurement of light: Errors in broad band photometry, Building Res. J. 2(1), 25-30. 4. Walsh, J. W. T. 1958. Photometry. 3rd ed. London: Constable. 5. Ohno, Yoshihiro. 1997. Photometric calibrations, NIST Special publication 250-237. Washington: National Institute of Standards and Technology. 6. Gaertner, A. A. 1994. Photometric and radiometric quantities. In Course on photometry, radiometry, colorimetry, NRC 38643. Ottawa, ON: National Research Council, Institute for National Measurement Standards. 7. Projector, T. E. 1957. Effective intensity of flashing lights. Illum. Eng. 52(12):630-640. 8. Douglas, C. A. 1957. Computation of the effective intensity of flashing lights. Illum. Eng. 52(12):641-646. 9. Lash, J. D., and G. F. Prideaux. 1943. Visibility of signal lights. Illum. Eng. 38(9):481-492. 10. Preston, J. S. 1941. Note on the photoelectric measurement of the average intensity of fluctuating light sources. J. Sci. Inst. 18(4):57-59. 11. Schuil, A. E. 1940. The effect of flash frequency on the apparent intensity of flashing lights having constant flash duration. Trans. Illum. Eng. Soc. (London) 35:117. 12. Neeland, G. K., M. K. Laufer, and W. R. Schaub. 1938. Measurement of the equivalent luminous intensity of rotating beacons. J. Opt. Soc. Am. 28(8):280-285. 13. Blondel, A., and J. Rey. 1912. The perception of lights of short duration at their range limits. Trans. Illum. Eng. Soc. 7(8):625-662. 14. MacGregor-Morris, J. T., and R. M. Billington. 1936. The selenium rectifier photo-electric cell: Its characteristics and response to intermittent illumination. J. Inst. Elec. Eng. (London) 77(478):435-438. 15. Zworykin, V. K., and E. G. Ramberg. 1949. Photoelectricity and its application. New York: Wiley, 211. 16. Gleason, P. R. 1934. Failure of Talbot's Law for barrier-layer photocells [abstract]. Phys. Rev., 2nd series, 45(10):745. 17. Lange, B. 1938. Photoelements and their application. Translated by A. St. John. New York: Reinhold, 151. 18. Commission Internationale de l'Éclairage. 1987. Methods of characterizing illuminance meters and luminance meters: Performance, characteristics and specifications, CIE publication No. 69. Vienna: Bureau Central de la CIE.

19. Rea, M. S., and I. G. Jeffrey. 1990. A new luminance and image analysis system for lighting and vision: I. Equipment and calibration. J. Illum. Eng. Soc. 19(1):64-72. 20. Lewin, I., R. Laird, and J. Young. 1992. Video photometry for quality control. Light. Des. Appl. 22(1):16-20. 21. American Society for Testing Materials. 1996. ASTM standards on color and appearance measurements. 5th ed. Philadelphia: ASTM. 22. IES. Committee on Testing Procedures. Subcommittee on General Photometry. 1974. IES approved method of reflectometry. J. Illum. Eng. Soc. 3(2):167-169. 23. Baumgartner, G. R. 1937. A light-sensitive cell reflectometer. Gen. Elec. Rev. 40(11):525-527. 24. Taylor, A. H. 1935. Errors in reflectometry. J. Opt. Soc. Am. 25(2):51-56. 25. Dows, C. L., and G. R. Baumgartner. 1935. Two photo-voltaic cell photometers for measurement of light distribution. Trans. Illum. Eng. Soc. 30(6):476-491. 26. Permanent gloss standards. 1950. Illum. Eng. 45(2):101. 27. Spencer, D. E., and S. M. Gray. 1960. On the foundations of goniophotometry. Illum. Eng. 55(4):228-234. 28. Nimeroff, I. 1952. Analysis of goniophotometric reflection curves. J. Res. Natl. Bur. Stand. 48(6):441-448. 29. American Society for Testing and Materials. 1996. American national standard practice for goniophotometry of objects and materials, ANSI/ASTM E167-96. Philadelphia: ASTM. 30. IES. Committee on Testing Procedures. Photometry of Light Sources Subcommittee. 1983. IES guide to spectroradiometric measurements. J. Illum. Eng. Soc. 12(3):136-140. 31. Cunningham, R. C. 1974. Silicon photodiode or photomultiplier tube? Electro-Opt. Sys. Des. 6(8):21-26. 32. Bode, D. E. 1971. Optical detectors. Chapter 5 in Handbook of Lasers, edited by R. J. Pressley. Cleveland, OH: Chemical Rubber Company. 33. Bode, D. E. 1980. Infrared detectors. Volume 6 in Applied optics and optical engineering, edited by R. Kingslake and B. J. Thompson. New York: Academic Press. 34. Weekes, F. 1977. Photon counting: Notes on a basic system. Electro-Opt. Sys. Des. 9(6):30-34. 35. Morton, G. A. 1968. Photon counting. Appl. Opt. 7(1):1-10. 36. McCulloch, J. H., and H. McCulloch. 1967. Floodlight photometry without special photometer and without tipping luminaire: A computer application. Illum. Eng. 42(4):243-245. 37. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1989. IES approved method for photometric testing of floodlights using incandescent filament or discharge lamps, IES LM-35-1989. New York: Illuminating Engineering Society of North America. 38. Rosa, E. B., and A. H. Taylor. 1922. Theory, construction, and use of the photometric integrating sphere: Paper No. 447. Sci. Pap. Bur. Stand. 18:281-325. 39. Buckley, H. 1946. The effect of non-uniform reflectance of the interior surface of spherical photometric integrators. Trans. Illum. Eng. Soc. (London) 41:167. 40. Hardy, A. C., and O. W. Pineo. 1931. The errors due to the finite size of holes and sample in integrating spheres. J. Opt. Soc. Am. 21(8):502-506. 41. Gabriel, M. H., C. F. Koenig, and E. S. Steeb. 1951. Photometry: Parts I and II. Gen. Elec. Rev. 54(9):30-37, 54(10): 23-29. 42. DeCusatis, C. 1997. Handbook of applied photometry. New York: AIP Press. 43. Commission Internationale de l'Éclairage. 1973. Procedures for the measurement of luminaire flux of discharge lamps and for their calibration as working standards, CIE publication no. 25. Paris: Bureau Central de la CIE. 44. Weaver, K. S., and B. E. Shackleford. 1923. The regular icosahedron as a substitute for the Ulbricht sphere. Trans. Illum. Eng. Soc. 18(3):290-304. 45. IES Committee on Testing Procedures for Illumination Characteristics. 1955. IES general guide to photometry. Illum. Eng. 50(4):201-210. 46. Stephenson, H. F. 1952. The equipment and functions of an illumination laboratory. Trans. Illum. Eng. Soc. (London) 17 (1):1-29. 47. IES. Committee on Testing Procedures. Subcommittee on Practical Guide to Photometry. 1971. IES practical guide to photometry. J. Illum. Eng. Soc. 1(1):73-96. 48. Levin, R. E. 1982. The photometric connection: Parts 1-4. Light. Des. Appl. 12(9):28-35, 12(10):60-63, 12(11):42-47, 12(12):16-18. 49. International Organization for Standardization and International Electrotechnical Commission. 1990. General requirements for the competence of calibration and testing laboratories, ISO/IEC Guide 25. Geneva: ISO. 50. IES. Committee on Testing Procedures. Subcommittee of Outdoor Luminaires. 1989. IES guide for the selection, care and use of electrical instruments in the photometric laboratory,. IES LM-28-1989. New York: Illuminating Engineering Society of North America. 51. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1991. IES guide to lamp seasoning, IES LM-54-1991. New York: Illuminating Engineering Society of North America. 52. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1988. IES approved method for the electrical and photometric measurements of fluorescent lamps, IES LM-9-1988. New York: Illuminating Engineering Society of North America. 53. IES. Committee on Testing Procedures. Photometry of Light Sources Subcommittee. 1993. IES approved method for the electrical and photometric measurements of high intensity discharge, IES LM-51-1993. New York: Illuminating Engineering Society of North America.

54. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1991. IES approved method of life testing of low pressure sodium lamps, IES LM-60-1991. New York: Illuminating Engineering Society of North America. 55. American National Standards Institute. 1998. Electric lamp ballasts--Line frequency fluorescent lamp ballasts, ANSI C82.1-1998. New York: ANSI. 56. American National Standards Institute. 1993. High frequency fluorescent lamp ballasts, ANSI C82.11-1993. New York: ANSI. 57. American National Standards Institute. 1995. Fluorescent lamp reference ballasts, ANSI C82.3-1995. New York: ANSI. 58. American National Standards Institute. 1992. Ballasts for high-intensity discharge and low pressure sodium lamps (multiple supply type), ANSI C82.4-1992. New York: ANSI. 59. American National Standards Institute.1995. Reference ballasts for high-intensity discharge and low pressure sodium lamps, ANSI 82.5-1995. New York: ANSI. 60. American National Standards Institute. 1996. Reference ballasts for high intensity discharge lamps--Methods of measurement, ANSI C82.6-1996. New York: ANSI. 61. American National Standards Institute. 1995. Fluorescent lamp ballasts: Methods of measurement, ANSI C82.2-1995. New York: ANSI. 62. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1990. IES approved method for electrical and photometric measurements of general service incandescent filament lamps, IES LM-45-91. New York: Illuminating Engineering Society. 63. Teele, R. P. 1930. Gas-filled lamps as photometric standards. Trans. Illum. Eng. Soc. 25(1):78-96. 64. Knowles-Middleton, W. E., and E. G. Mayo. 1951. Variation in the horizontal distribution of light from candlepower standards. J. Opt. Soc. Am. 41(8):513-516. 65. Winch, G. T. 1956. Recent developments in photometry and colorimetry. Trans. Illum. Eng. Soc. (London) 21(5):91-116. Also see reference 3. 66. Winch, G. T. 1946. Photometry and colorimetry of fluorescent and other discharge lamps. Trans. Illum. Eng. Soc. (London) 21:107. 67. Voogd, J. 1939. Physical photometry. Philips Tech. Rev. 4(9):260-266. 68. Winch, G. T. 1949. The measurement of light and colour. Proc. Inst. Elec. Eng. (London) 96(2):452-470. 69. Franck, K., and R. L. Smith. 1954. A photometric laboratory for today's light sources. Illum. Eng. 49(6):287-291. 70. Baumgartner, G. R. 1950. New semi-automatic distribution photometer and simplified calculation of light flux. Illum. Eng. 45(4):253-261. 71. Baumgartner, G. R. 1941. Practical photometry of fluorescent lamps and reflectors. Illum. Eng. 36(10):1340-1353. 72. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1982. IES approved method for photometric measuring and reporting tests on reflector type lamps. J. Illum. Eng. Soc. 11(3):130-134. 73. Eastman, A. A., and J. H. Campbell. 1952. Stroboscopic and flicker effects from fluorescent lamps. Illum. Eng. 47(1): 27-35. 74. Spears, G. R. 1974. Spectroradiometry photometry. J. Illum. Eng. Soc. 3(3):229-233. 75. Elby, J. E. 1970. A computer based spectroradiometer system. Appl. Opt. 9(4):888-894. 76. Lewin, I., G. A. Baker, and M. T. Baker. 1979. Developments in high speed photometry and spectroradiometry. J. Illum. Eng. Soc. 8(4):214-219. 77. National Bureau of Standards. 1963. Comparing materials or products with respect to average performance. Chapter 3 in Experimental statistics, edited by M. Gibbons. Handbook, 91. Washington: U.S. G.P.O. 78. American Society for Testing Materials. 1958. Standard practice for probability sampling of materials, ASTM E10558. Philadelphia: ASTM. 79. National Bureau of Standards. 1963. Characterizing the measured performance of a material, product or process. Chapter 2 in Experimental statistics, edited by M. Gibbons, Handbook, 91. Washington: U.S. G.P.O. 80. Lewinson, L. J. 1916. The interpretation of forced life tests of incandescent electric lamps. Trans. Illum. Eng. Soc. 11(8): 815-835. 81. Millar, P. S., and L. J. Lewinson. 1911. The evaluation of lamp life. Trans. Illum. Eng. Soc. 6(8):774-781. 82. Purcell, W. R. 1949. Saving time in testing life. Elec. Eng. 68(7):617-620. 83. IES. Committee on Testing Procedures. 1979. IES approved method for life testing of general lighting incandescent filament lamps. J. Illum. Eng. Soc. 8(3):152-154. 84. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1987. IES approved method for life performance testing of fluorescent lamps, IES LM-401987. New York: Illuminating Engineering Society of North America. 85. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1995. IES approved method for life testing of high intensity discharge (HID) lamps, IES LM47-1995. New York: Illuminating Engineering Society of North America. 86. IES. Committee on Testing Procedures. Subcommittee on Photometry of Light Sources. 1998. IES approved method for the electrical and photometric measurements of low pressure sodium lamps, IES LM-59-1998. New York: Illuminating Engineering Society of North America. 87. IES. Committee on Testing Procedures. Subcommittee on Photometry of Indoor Luminaires. 1998. Approved method for photometric testing of indoor fluorescent luminaires, IES

LM-41-1998. New York: Illuminating Engineering Society of North America. 88. IES. Committee on Testing Procedures. Subcommittee on Photometry of Indoor Luminaires. 1998. IESNA approved method for photometric testing of indoor luminaires using high intensity discharge or incandescent filament lamps, IES LM-46-1998. New York: Illuminating Engineering Society of North America. 89. IES. Committee on Testing Procedures. 1972. IES guide for reporting general lighting equipment engineering data. J. Illum. Eng. Soc. 1(2):175-180. 90. IES. Committee on Testing Procedures. 1976. Addendum to IES guide for reporting general lighting equipment engineering data. J. Illum. Eng. Soc. 5(4):243. 91. IES. Committee on Testing Procedures. Subcommittee on Photometry of Indoor Luminaires. 1972. Determination of average luminance of luminaires. J. Illum. Eng. Soc. 1(2): 181-184. 92. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1996. IES approved guide for identifying operating factors for installed high intensity discharge (HID) luminaires, IES LM-61-1996. New York: Illuminating Engineering Society of North America. 93. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1984. IES guide for photometric testing of searchlights. J. Illum. Eng. Soc. 13(4): 372-380. 94. National Electrical Manufacturers Association. 1973. Outdoor floodlighting equipment, NEMA FA1-1973 (R1979). Washington: NEMA.. 95. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1989. IES approved method for photometric testing of floodlights using incandescent filament or discharge lamps, IES LM-35-1989. New York: Illuminating Engineering Society of North America. 96. Joint IES-SMPTE Committee on Equipment Performance Ratings. 1958. Recommended practice for reporting photometric performance of incandescent filament lighting units used in theatre and television production. Illum. Eng. 53(9):516-520. 97. Commission Internationale de l'Éclairage. 1979. Photometry of floodlights. CIE Publication No. 43. Prepared by CIE Technical Committee TC 2.4. Paris: Bureau Central de la CIE. 98. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1995. IES approved method for photometric testing of roadway luminaires using incandescent filament and high intensity discharge lamps, IES LM-31-1995. New York: Illuminating Engineering Society of North America. 99. National Bureau of Standards. 1963. Photometry of projectors at the National Bureau of Standards. NBS Tech. Note 198. Gaithersburg, MD: National Bureau of Standards. 100. Johnson, J. 1962. Zero-length searchlight photometry system. Illum. Eng. 57(3):187-194. 101. Stephenson, H. F. 1952. The equipment and functions of an illumination laboratory. Trans. Illum. Eng. Soc. (London) 17(1):1-29. 102. U.S. Civil Aeronautics Authority. 1944. Construction of a goniometer for use in determining the candlepower characteristics of beacons. CAA Technical Development Report No. 39. Prepared by F. C. Breckenridge and T. H. Projector. Washington: U.S. G. P. O. 103. Projector, T. H. 1953. Versatile goniometer for projection photometry. Illum. Eng. 48(4):192-196. 104. Frederiksen, E. 1967. Unidirectional-sensitive photometer. Light. 60(2):46-48. 105. IES. Committee on Testing Procedures. Subcommittee on Guide for Measurement of Photometric Brightness. 1961. IES guide for measurement of photometric brightness (luminance). Illum. Eng. 56(7):457-462. 106. Lindsay, E. A. 1944. Brightness of cylindrical fluorescent sources. Illum. Eng. 39(1):23-30. 107. Horton, G. A. 1950. Modern photometry of fluorescent luminaires. Illum. Eng. 45(7):458-467. 108. IES. Committee on Testing Procedures. Subcommittee on Photometry of Indoor Luminaires. 1978. IES approved guide for the photometric and thermal testing of air cooled heat transfer luminaires. J. Illum. Eng. Soc. 8(1):57-62. 109. Carter, D. J., R. C. Sexton, and M. S. Miller. 1989. Field measurement of illuminance. Light Res. Tech. 21(1) 29-35. 110. Joint Lighting Survey Committee of the Illuminating Engineering Society and the U.S. Public Health Service. 1963. How to make a lighting survey. Illum. Eng. 57(2): 87-100. 111. Ouellette, M. J., B. W. Transley, and I. Pasini. 1993. The dilemma of emergency lighting: Theory versus reality. J. Illum. Eng. Soc. 22(1)113-121. 112. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1991. IES guide for photometric measurement of roadway lighting installations, IES LM-50-1991. New York: Illuminating Engineering Society of North America. 113. IES. Committee on Testing Procedures. Subcommittee on Photometry of Outdoor Luminaires. 1996. Photometric measurements of area and sports lighting installations, IES LM-596. New York: Illuminating Engineering Society of North America. 114. IES. Committee on Testing Procedures. Subcommittee on Photometry for Outdoor Luminaires. 1998. IES guide for photometric measurements of roadway sign installations, IES LM-52-1998. New York: Illuminating Engineering Society of North America. 115. IES. Committee on Sports and Recreational Areas, and Committee on Theatre, Television and Film Lighting. 1969. Interim Report: Design criteria for lighting of sports events for color television broadcasting. Illum. Eng. 64(3):191-195.

3 Vision and Perception . . . [S]eeing, regarded as a supply for the primary wants of life is in its own right the superior sense --Aristotle (384-322 BC)

INTRODUCTION Vision depends on light. Lighting should provide visual conditions in which people can function effectively, efficiently, and comfortably. To predict human behavior as a function of the lighting conditions, it is important to understand the physical, physiological, and perceptual characteristics of the visual system. This chapter highlights some of the basic relationships between light and vision. It provides some fundamental data that the illuminating engineer may find useful, and calls attention to the factors that need to be considered when designing lighting for visual performance and comfort.

VISUAL SYSTEM STRUCTURE The visual system is an image processing system. It involves the eye and brain working together to interpret the visual environment (Figures 3-1 and 3-2). The optical elements of the eye form an image of the world on the retina. At the retina, photons of light are absorbed by the photoreceptors and converted to electrical signals. These signals are transmitted by the optic nerve to the lateral geniculate nucleus (LGN) and then to the visual cortex for visual processing. In addition to the neural pathways from the eye to the visual cortex, there are a number of other pathways leaving the optic nerve shortly after it exits the eye that control pupil size, eye movements, and circadian rhythms.

The Eye The structure of the eye can be divided into three distinct parts: the ocularmotor components (the eye muscles), the optical components (the cornea, crystalline lens, pupil, and intraocular humors), and the neurological components (the retina and optic nerve). Ocularmotor Components. The ocularmotor components of the eye consist of three pairs of muscles (Figure 3-3). These muscles position the lines of sight of the two eyes so that they are both pointed towards the same object of regard (Figures 3-4 and 3-5). The line of sight of the eye passes through the part of the retina used for discriminating fine detail, the fovea. If the image of a target does not fall on the fovea, the resolution of target detail will be reduced. Additionally, if the foveas of both eyes are not aimed at the same target, the target may be seen as double (diplopia). Eye movements can take several different forms.1 Among the more important are: 1. Saccades. High-velocity movements, usually generated to move the line of sight from one target to another, are called saccades. Velocities may range up to 1000 degrees per second, depending upon the distance moved. Saccadic eye movements have a latency of 150 to 200 ms, which limits how frequently the line of sight can be moved in a given time period; approximately five movements per second is the maximum. Visual functions are substantially limited during saccadic movements. Eye movements during reading characterize a series of alternate fixations and saccades, along a row of print.

2. Pursuit. Smooth eye movements called pursuits are used to follow a smoothly moving target after a saccade has been used to bring the retinal image of the target onto the fovea. The pursuit system cannot follow smoothly moving targets at high velocities, nor can it follow slowly but erratically moving targets. If the eye cannot follow the target, resolution of target details decreases because the target's retinal image is no longer on the fovea. To catch up, binocular pursuit and jump movements are made, which are referred to as version movements when they involve objects in a frontal plane. For these movements, the two eyes make equal movements in the same direction, so there is no change in their angle of convergence (Figure 3-4). 3. Vergence movements. Movements of the two eyes that keep the primary lines of sight converged on a target or that may be used to switch fixation from a target at one distance to a new target at a different distance are called vergence movements (Figure 3-5). These can occur as a jump movement or can smoothly follow a target moving in a fore-and-aft direction. Both types of movement involve a change in the angle between the eyes. When the primary lines of sight drift apart so that they fail to converge at the intended point of fixation, vergence movements play a major role in making the eyes see the target.

Figure 3-1. A horizontal section through the eye. The approximate length from the cornea to the retina is 24 mm. The thickness of the choroid is about 0.05 mm and the sclera 1.0 mm. Optical Components. The function of the optical components of the eye is to form an image of the target on the retina. For this to occur, light has to be transmitted through the eye without excessive absorption and scattering, and the image of the target has to be focused on the retina (Figure 3-1).

Figure 3-2. A schematic diagram of the structure of the visual system. Used, by permission, from R. Sekuler and R. Blake, Perception. © 1994. McGraw-Hill. The transmittance of the eye varies with wavelength and with age.2 In young and old eyes, the cornea absorbs most of the incident radiation shorter than 300 nm. In contrast, the human crystalline lens gradually develops a yellow pigmentation as it ages. This pigmentation attenuates the total transmission of radiant energy to the retina, especially in the shorter-wavelength portion of the visible and UV spectrum (Figure 36).3 As shown in more recent work (see Chapter 5, Nonvisual Effects of Optical Radiation, and Figure 5-4), newborn human lenses transmit UV energy.4 This transmission is greatly reduced but not entirely lost by early adulthood. Later in adulthood the UV transmission is entirely lost, and there are also significant reductions in transmission of short-wavelength portions of the visible spectrum. Accordingly, the retina receives radiation in the range from 380 to 950 nm with limited attenuation. Beyond 950 nm, transmittance is variable, with major absorption in the infrared (IR) water bands. Very little IR radiation beyond 1400 nm reaches the retina.

Figure 3-3. An eye and the extraocular muscles used to move it. Used, by permission, from R. Sekuler and R. Blake, Perception. © 1994. McGraw-Hill. In the visible region of the spectrum, the optics of the eye transmit more light at long wavelengths (the red end) than at short wavelengths (the blue end), a tendency that is enhanced at the fovea by the additional short wavelength absorption of the macula lutea, a yellow filter that lies immediately above the fovea and parafovea. On average, some 70 to 85% of the visible spectrum reaches the retina in young eyes.4 As one ages, there is a general reduction in the transmittance at all wavelengths combined with a marked reduction (greater than 4 times) in short-wavelength transmittance, due primarily to thickening and yellowing of the

crystalline lens (Figure 3-7).5 While absorption of light reduces the magnitude of the stimulus to the visual system, it does not degrade the quality of the retinal image, that is, it does not blur the retinal image nor reduce its luminance contrast. Such degradation occurs when light is scattered in the eye or additional light is generated within the eye. Scattering within the eye is primarily large-particle scattering, which is not wavelength dependent. In young eyes, some 25% of the scattered light is produced by the cornea,6 another 25% by the fundus7-9 (see "Neurological and Supportive Components" below), and the rest by the lens and the vitreous humor. The aqueous humor causes little scattering, if any. The amount of scattered light in the eye increases with age. Consequently, older eyes are more susceptible to disability glare, as discussed later. Almost all of the increased scattering with age is due to changes in the lens.10 The quality of the retinal image can also be reduced by light generation within the eye, caused by fluorescence in the lens. This phenomenon occurs primarily in the elderly and is produced by absorption of short wavelength visible and ultraviolet radiation in the lens which is then re-emitted at longer wavelengths to which the visual system is more sensitive.11

Figure 3-4. In a version movement, as the target moves from point O to point O' the angle between the eyes remains constant. There are three optical components involved in the ability of the eye to refract or to focus an image on the retina. The first is the thin film of tears on the cornea. This film is important because it cleans the surface of the eye, starts the optical refraction (light bending) process necessary for focusing objects, and smoothes out small imperfections in the surface of the cornea. The second optical component is the cornea. This covers the transparent anterior one-fifth of the eyeball (Figure 3-1). With the tear layer, it forms the major refracting component of the eye and gives the eye about 70% of its refractive power. The crystalline lens provides most of the remaining 30% of the refracting power. The ciliary muscles have the ability to change the curvature of the lens and thereby adjust the power of the eye's optical system, when needed, in response to changing target distances or certain types of refractive errors; this change in power is called accommodation.

Figure 3-5. In a vergence movement, as the target moves from point O to point O', the angle between the eyes changes. Accommodation is always a response to an image of the target located on or near the fovea rather than in the periphery. It is used to bring a defocused image into focus or to change focus from one target to another at a different distance. It may be gradually changed to keep in focus a target that is moving in a fore-and-aft direction. Any condition, either physical or physiological, that handicaps the fovea, such as a low light level, will adversely affect accommodative ability. Blurred vision and eyestrain can be consequences of limited accommodative ability.12

Figure 3-6. Spectral transmission properties of the human ocular media. The solid curves refer to the total light transmittance through the medium. The dashed curves refer to the direct, unscattered components only. The difference between the solid and dashed curve at each wavelength indicates the amount of light that is scattered by transmission through the medium. Where more than one dashed curve is shown the lower ones are for older eyes.

When there is no stimulus for accommodation, as in complete darkness or in a uniform luminance visual field such as occurs in a dense fog, the accommodation system typically accommodates to approximately one meter away.13 Neurological and Supportive Components. The posterior 80% of the eye is enclosed by three layers of tissue (Figure 3-1). Collectively, they protect and nourish the eye and transduce light into electrical signals:

1. The sclera. The outermost covering of the globe, which is continuous with the cornea, protects the eye's contents and defines its shape. 2. The choroid. A highly vascular tissue that contains the blood supply to much of the eye. 3. The retina. The innermost layer, which converts radiant energy into electrical signals that are sent to the brain. Together, the choroid and the retina constitute the fundus.

Figure 3-7. The optical density of the human lens at 490nm as a function of age (optical density D = log (1/t) where t = total transmittance). Photoreceptors. The retina contains two main classes of light-sensitive receptors, rods and cones, which are differentiated by their morphology and by the spectral sensitivity of the photopigments which they contain (Figure 3-8).

Figure 3-8. A simplified diagram of the connections among the neural elements in the retina. The regions where the cells are contiguous are synapses. The direction of incident light is from the bottom of this diagram. Rods, which are absent in the fovea, increase in number to a maximum at about 20° of eccentricity and then gradually decrease towards the edges of the retina (Figure 3-9). All rods contain the same photopigment (rhodopsin), which has a peak spectral sensitivity at approximately 507 nm (Figure 3-10).

Figure 3-9. The distribution of rod and cone photoreceptors across the retina. The 0° point represents the fovea.

Figure 3-10. The CIE Standard Photopic and Scotopic Observers, representing the relative spectral sensitivity of the cone and rod photoreceptors, respectively. Cones are divided into three known classes, each characterized by the photopigment that it contains: erythrolabe, chlorolabe, or cyanolabe (also known as L-type, M-type, and S-type or long-, middle-, and short-wavelength type) (Figure 3-11). Cones are concentrated in the fovea, although there are cones in all parts of the retina (Figure 3-9). All three cone types acting together have a peak spectral sensitivity at approximately 555 nm (Figure 3-10). The different photopigments in the cones make color discrimination possible.14 Receptive Fields. Photoreceptors do not send their information directly to the brain, but rather to several other cells in the retina, which in turn send them to ganglion cells, whose terminal axons constitute the optic nerve (Figure 3-8). In this way, light received by a number of receptors is "pooled" to provide a signal strong enough to stimulate a ganglion cell. The area of retina that stimulates a ganglion cell is called a receptive field. Although photoreceptors are the primary transducers of light into electrical signals, the receptive fields begin the image processing, which enables the visual system to interpret the visual environment.

Figure 3-11. The relative spectral sensitivity curves of the three cone photoreceptors: long (L), middle (M), and short (S). Ganglion cell receptive fields are comprised of two distinct, juxtaposed areas: a circular center and an annular surround. These two areas receive signals from different, individual bipolar cells (Figure 3-8), which have received information from different photoreceptors. In the fovea, the center area of the receptive field receives neural signals from a single bipolar cell, which itself receives signals from a single cone photoreceptor. At greater eccentricities from the fovea, receptive field centers are larger because they receive input from many photoreceptors, both rods and cones, through the bipolar layer. Receptive field surrounds commonly receive input from several bipolar cells, which are fed not only by direct links to photoreceptors but also from special cells in the retina which laterally connect other bipolar cells. Some of the lateral connections are illustrated in the bipolar layer of Figure 3-8. These facts reflect the tradeoff in the retina between fine spatial resolution and high sensitivity to light. Large receptive fields, like those found in the periphery, can gather very few photons and sum them to produce a neural signal for "light." Every photon captured within a receptive field produces the same neural response; hence, the location of each photon capture within a receptive field cannot be spatially segregated. Very small receptive fields, like those in the fovea, are needed to precisely locate objects on the retina. Thus, the fovea has excellent spatial resolution but low sensitivity to light, whereas the periphery has high sensitivity to light and poor spatial resolution. A very important aspect of this center-surround organization is the ability to enhance the contrast of images at the boundary. The eye is a very poor optical system fraught with many types of aberrations and imperfections (see "Focusing Problems"). To overcome its severe optical limitation, the visual system has developed the center-surround receptive field organization to provide a simple, yet elegant, image enhancement system. The center and surround areas produce opposite neural polarities in the ganglion cell when stimulated by light. Light striking the center of the receptive field will increase the neural firing rate of the ganglion cell, whereas the same light striking the surround will decrease the firing rate. Light striking both the center and surround will produce an intermediary response because the excitation produced by the light stimulating the center area will be counteracted by the inhibitory effect of light stimulating the surround. Consider then an image of a white disc on a black background focused and precisely covering the center of a receptive field. This configuration will produce the maximum excitation to the ganglion cell because the

center is maximally stimulated and the surround is minimally stimulated. The net effect is the strongest possible signal from the ganglion cell and, therefore, the highest possible contrast between the white center and black surround. Larger or smaller discs imaged at the same location will have smaller impact on the ganglion cell; a larger disc will increase the inhibitory response of the surround, and a smaller disc will not stimulate the center as strongly because it does not completely cover the center of the receptive field. With a little thought, the magnitude of contrast enhancement for other image shapes can be surmised on the basis of how their edges are positioned within the receptive field. The common result, however, is that the contrast of juxtaposed light and dark areas of an image will be enhanced if they are focused at the boundary between a receptive field center and its surround. At any location in the retina there is an equal proportion of receptive fields with excitatory centers and inhibitory centers. Therefore, the contrast of both black discs and white discs will be enhanced. Also at any location in the retina, there is a distribution of receptive field sizes. This enables the retina to enhance the contrast of variously sized images. Nevertheless, there is an optimum target size for any retinal location; smaller images are seen best in the fovea, and larger ones are seen best in the periphery. It is important to note that receptive field sizes are not constant, but rather they change size with light level. As light level increases, receptive field sizes increase as they collect signals through their lateral connections from more distant locations in the retina. In effect, this greater inhibition from the receptive field surround makes the center of the receptive field functionally smaller. Indeed, the center of a receptive field in the fovea can become smaller than the diameter of single cone at high light levels. This reduction in the size of the receptive field center enables us to improve acuity as light level is increased (Figure 3-27). Color vision also depends on this organization. Consider a receptive field where L-cones exclusively populate an excitatory center and M-cones populate an inhibitory surround. Consider now two colored lights that cover the entire receptive field. If the L-cones are stimulated more than the M-cones for one of the colored lights this ganglion cell will signal "red" by increasing its neural firing rate. If the other light produces a greater response in the M-cones, the ganglion cell will signal "green" because its firing rate decreases. Because of this receptive field polarity, red and green hues are perceptual opposites; this is why one cannot see, for example, a "reddish green" light. Yellow and blue are also opposed through ganglion cell receptive field organization. Yellow is created by the sum of the input of the L and M cones in opposition to blue, which is produced by the input from the S cone. There are also some spatial implications of this receptive field organization. For example, very small images in the center of the fovea cannot produce perceptions of yellow or blue because there are no S-cones in the center of the fovea. Further information on color vision may be found throughout this chapter and in Chapter 4, Color. Neural Pathways. Electrical signals from the receptive fields in the retina are transmitted over the optic nerve. Approximately 20% of optic nerve fibers project to the superior colliculus, and 80% to the lateral geniculate nucleus (LGN) and on to the visual cortex (Figure 3-2). At the optic chiasm, the fibers from each eye divide into two sets; one set remains on the same side of the head as the eye, and the other set crosses over to the other side. The result is two optic tracts, both of which contain nerve fibers from both eyes; one tract transmits the signal from the left side of both eyes to the left side of the visual cortex, and the other transmits the signals from the right side of both eyes to the right side of the visual cortex (Figure 3-2). The superior colliculus is a phylogenetically older part of the brain and in humans is involved in controlling eye movements. Because the superior colliculus also receives signals from the ears, it is believed that its role is directing eye and head movements towards targets located away from the point of fixation. It is not involved in the detailed image processing. The LGN contains an orderly representation of the retina. Investigations of its functioning have revealed that it continues processing the retinal image by sorting the information it contains into distinct categories. This is accomplished with two channels of information flow, the magnocellular channel and the parvocellular channel. The magnocellular channel transmits primarily temporal information and is dominant in the periphery of the retina. The parvocellular channel transmits primarily spatial information and is dominant in the fovea. This pattern is consistent with how the periphery of the retina identifies changes in

the visual environment and the fovea determines the nature of those changes. The visual cortex takes the information sorted by the LGN and refines and interprets it in terms of past experience. Approximately 80% of the visual cortex is assigned to analyze and interpret the central 10° of the visual field. This pattern of assignment is called cortical magnification. Although the components involved in transforming patterns of photons of light into visual perceptions have been discussed separately, it is important to appreciate that there is considerable interaction between them. An example of such interactions is the system of color vision (see Chapter 4, Color, and "Color Discrimination" in this chapter). The ability to discriminate among wavelengths of light is due to a combination of photochemical and neurological processes. Signals from the three cone types are coded in the retina and the lateral geniculate body into chromatic and achromatic information. As a first-order model of color and brightness perception, chromatic information is a result of a subtraction of photoreceptor signals, while achromatic information is a result of the addition of photoreceptor signals. However, many experiments using various testing procedures and stimuli demonstrate that this is an oversimplification of how the visual system processes light signals. Chromatic, achromatic, spatial, and temporal information are combined in complicated ways to give final perceptions of light and color. For example, equal-luminance colored lamps may have different apparent brightnesses because of the interaction between achromatic and chromatic channels.15 Figure 3-12 is a proposed model of how the visual system combines the information from these various channels to produce human perceptions.16

Dark and Light Adaptation For the visual system to be able to function well, it has to be adapted to the prevailing light condition. The human visual system can process information over an enormous range of luminances (approximately 12 log units), but not all at once. To cope with the wide range of retinal illuminations to which it might be exposed, from a dark night to a sunlit beach, the visual system changes its sensitivity through a process called adaptation. Adaptation involves three distinct processes:

Figure 3-12. A proposed model for the neural connections in the visual system. Information from

the first stage photoreceptors (R,G,B) goes to mechanisms that sum or subtract input to give achromatic and chromatic information, respectively. Subsequent "cortical analysis" mechanisms receive multiple inputs from the second stage. Such a model attempts to qualitatively describe some of the nonlinearities in the visual system that have been discovered using stimuli that vary in several dimensions (spatial, temporal, chromatic and achromatic). The solid lines indicate wellestablished inputs, while the dashed lines are more speculative.

Figure 3-13. Pupil diameter for light-adapted (open circles) and dark-adapted (filled circles) conditions, plotted against the age of the observer. From R.A. Weale, The senescence of human vision, © 1992, used by permission of Oxford University Press. 1. Change in Pupil Size. The iris (Figure 3-1) constricts and dilates in response to increased and decreased levels of retinal illumination. Iris constriction has a shorter latency and is faster (approximately 0.3 s) than dilation (approximately 1.5 s).17 There are wide variations in pupil sizes among individuals and for any particular individual at different times for the same visual stimulus. Pupil size is influenced by emotions, such as fear or elation. Thus, for a given luminous stimulus, some uncertainty is associated with an individual's pupil size until it is measured. The typical range in pupil diameter for young people is from 3 mm for high retinal illuminances to 8 mm for low retinal illuminances.18 This change in pupil size in response to retinal illumination can only account for a 1.2 log unit change in sensitivity to light. Older people tend to have smaller pupils under comparable conditions (Figure 3-13). 2. Neural Adaptation. This is a fast (less than 200 ms) change in sensitivity produced by synaptic interactions in the visual system.19 Neural processes account for virtually all the transitory changes in sensitivity of the eye where cone photopigment bleaching has not yet taken place (discussed below), in other words, at luminance values commonly encountered in electrically lighted environments, below approximately 600 cd/m2. The facts that neural adaptation is fast, is operative at moderate light levels, and is effective over a luminance range of 2 to 3 log units explain why it is possible to look around most lit interiors without being conscious of being misadapted. 3. Photochemical Adaptation. The retinal photoreceptors contain four photopigments. When light is absorbed, the pigment breaks down into an unstable aldehyde of vitamin A and a protein (opsin) and gives off energy that generates electrical signals that are relayed to the brain and interpreted as light. In the dark, the pigment is regenerated and is again available to absorb light. The sensitivity of the eye to light is largely a function of the percentage of unbleached pigment. Under conditions of steady retinal irradiance, the concentration of photopigment is in equilibrium; when the retinal irradiance is changed, pigment is either bleached or regenerated

to reestablish equilibrium. Because the time required to accomplish the photochemical reactions is on the order of minutes, changes in the sensitivity often lag behind the stimulus changes. The cone system adapts much more rapidly than does the rod system; even after exposure to high irradiances, the cones achieve their maximum sensitivity in 10 to 12 min, while the rods require 60 min (or longer) to achieve their maximum sensitivity (Figure 3-14).20

Figure 3-14. The increase in sensitivity to light (decrease in threshold) as a function of time in the dark, after exposure to a bright light. Sensitivity is measured at a point 7° from the fovea. The two curves represent the extremes of the normal range of observers. (1 picolambert = 3.2 × 10−9 cd/m2) Exactly how long it takes to adapt to a change in retinal illumination depends on the magnitude of the change, the extent to which it involves different photoreceptors, and the direction of the change. For changes in retinal illumination of approximately 2 to 3 log units, neural adaptation is sufficient, so adaptation is in less than a second. For larger changes, photochemical adaptation is necessary. If the change in retinal illumination lies completely within the range of operation of the cone photoreceptors, a few minutes is sufficient for adaptation to occur. If the change in retinal illumination covers from cone photoreceptor operation to rod photoreceptor operation, tens of minutes can be required. As for the direction of change, once the photochemical processes are involved, changes to a higher retinal illuminance can be achieved much more rapidly than changes to a lower retinal illuminance. When the visual system is not completely adapted to the prevailing retinal illumination, its capabilities are limited.21 This state of changing adaptation is called transient adaptation. Transient adaptation is unlikely to be noticeable in interiors in normal conditions but can be significant where sudden changes from high to low retinal illumination occur, such as on entering a long road tunnel on a sunny day or in the event of a power failure in a windowless building.

Photopic, Scotopic, and Mesopic Vision This process of adaptation takes the visual system through three distinct operating states. 1. Photopic vision. This operating state of the visual system occurs at luminances higher than approximately 3 cd/m2. For these luminances, the retinal response is dominated by the cone photoreceptors. This means that color is perceived and fine detail can be resolved in the fovea. 2. Scotopic vision. This operating state of the visual system occurs at luminances less than approximately 0.001 cd/m2. For these luminances only the rod photoreceptors respond to stimulation, so the fovea of the retina is inoperative. There is no perception of color, and what

resolution of detail there is occurs in the periphery within a few degrees of the fovea. 3. Mesopic vision. This operating state of the visual system is intermediate between the photopic and scotopic states. In the mesopic state both cones and rod photoreceptors are active. As luminance declines through the mesopic region, the fovea, which contains only cone photoreceptors, slowly declines in absolute sensitivity without significant change in spectral sensitivity,22 until vision fails altogether as the scotopic state is reached. In the periphery, the rod photoreceptors gradually come to dominate the cone photoreceptors, resulting in gradual deterioration in color vision and resolution and a shift in spectral sensitivity to shorter wavelengths. The relevance of these different operating states for lighting practice varies. Scotopic vision is largely irrelevant to lighting practice. Nearly every lighting installation provides enough light to at least move the visual system into the mesopic state. Most interior lighting ensures the visual system is operating in the photopic state. Current practice in exterior lighting ensures the visual system operates near the boundary of the photopic and mesopic states. The spectral sensitivities of the visual system in the photopic and scotopic states have been defined by the Commission Internationale de l'Éclairage (CIE). Figure 3-10 shows the CIE Standard Photopic and Standard Scotopic Observers. These two luminous efficiency functions are used in the fundamental definition of light, to convert from radiometric quantities to photometric quantities (see Chapter 1, Light and Optics). The mesopic state has been extensively studied but has not been defined officially by the CIE, partly because of problems with additivity.23-26 Problems with additivity are to be expected for any system based on brightness because brightness perception uses the parvocellular channel, which combines both achromatic and chromatic responses in a complex way. Recently, an alternative approach to mesopic photometry has been proposed, based on measurements of reaction time.27 Because scotopic vision is irrelevant and the mesopic state has not officially been defined, virtually all photometric quantities used in lighting practice are measured using the CIE Standard Photopic Observer, even for exterior lighting where the visual system may be operating in the mesopic state. It should be realized that the use of the CIE Photopic Observer can give discrepancies between measured photometric quantities in a space and the perception of brightness in the space. The CIE Standard Photopic Observer is based on the relative amounts of power at each wavelength required to produce a criterion brightness response in a 2° foveal field of view. It is an average response derived from several different experimental techniques, including techniques using direct brightness judgments involving the parvocellular channel and techniques are based on flicker perception involving the magnocellular channel.8 Studies have shown that, even with the visual system operating in the photopic state, the CIE Standard Photopic Observer slightly underestimates the influence of the short wavelength region of the visible spectrum on brightness, even for a 2° field of view;28 and the underestimation is greater for a 10° field of view,8 because this larger field extends beyond the macula lutea. This discrepancy between photometric quantities and brightness perception is slight for light sources with a spectral content distributed over the whole visible spectrum. However, when light sources with a very discrete spectral content are compared, the limitations of the CIE Standard Photopic Observer can become important. This is particularly so for colored signal lights, where the brightness of the light is what matters.29

Individual Differences Although the visual systems of all people have the same basic structure, as in most living things, there are individual differences. For example, Figure 3-15 shows a wide variation in luminous efficiency for 52 individuals.30 Many of these differences are ignored when considering lighting for use by the general public, but some are sufficiently large and their effects so predictable that they need to be taken into account in some lighting applications. This is especially true when lighting for the aged and partially sighted, as discussed later (see "Aging and Partial Sight").

Figure 3-15. The range of luminous efficiency values for 52 observers. Focusing Problems. As discussed above in "Ocular Components," the eye adjusts its optical power to focus objects at different distances on the retina. This is possible for a wide range of distances when there is a match between the combined optical power of the cornea and lens and the dimensions of the eye. However, when there is a mismatch between the optical power and the distance from the lens to the retina, a sharp image cannot be formed on the retina. This blurred retinal image is called a refractive error. There are several different forms of refractive error. They are: 1. Myopia. The optical power is greater than necessary so objects at a distance are focused in front of the retina (Figure 3-16a). 2. Hyperopia. The optical power is less than necessary so objects at a distance are focused behind the retina (Figure 3-16b). 3. Astigmatism. The optical power is not equal in all planes so objects are focused in front of, behind and on the retina for different planes (Figure 3-16c). 4. Presbyopia. The adjustment of optical power is limited. Typically, near objects are focused behind the retina (Figure 3-16d).

Figure 3-16. The relationship between the image of a point object and the retina in common refractive errors. (a) In myopia, the image forms in front of the retina. (b) In hyperopia, the image forms behind the retina. (c) In astigmatism, multiple foci are formed due to different optical powers occurring in the various meridians of the eye. (d) In presbyopia, accommodation is sufficiently limited that near objects focus behind the retina. Most of these refractive errors can be corrected by the use of spectacles or contact lenses, although even when the eye is perfectly corrected for refractive errors, a residual blur can remain due to spherical and chromatic aberrations. 1. Spherical aberration. Light rays that enter through the periphery of the cornea are refracted more than those that enter through the central zones (Figure 3-17). Thus, light in the retinal image is partially redistributed over a larger retinal area than would be the case in an aberration-free system. The amount and type of spherical aberration varies with the state of accommodation (Figure 3-18). 2. Chromatic aberration. Shorter wavelengths are refracted more than longer wavelengths (Figure 3-19). As in spherical aberration, the results of the different foci cause blur (Figure 320).

Figure 3-17. Spherical aberration: marginal rays (F'm) are focused in front of rays entering the eye near the center of the pupil (F'p).

These aberrations (and others) are mainly of theoretical interest. They are partially compensated by the image processing of the visual system and usually can be neglected in practical lighting design. They may, however, be important in certain specialized applications, such as work under reduced illuminances where pupil sizes can be large. Abnormal Color Vision. Approximately 8% of males and 0.2% of females have some form of abnormal color vision. Abnormal color vision occurs because of abnormal photoreceptor photopigments. The reason for the preponderance of males is that abnormal color vision is due to a genetic difference on the Xchromosome. Males have only one X-chromosome, but females have two, and for a female to have abnormal color vision, both X-chromosomes must have the same abnormal gene.

Figure 3-18. Spherical aberration: the amount of spherical aberration (in diopters, D) is on the horizontal axis (positive when undercorrected) and the distance from the achromatic axis is on the vertical axis. The solid line corresponds to the unaccommodated eye; the dashed line corresponds to 1.5 D accommodation; the dotted line corresponds to 3.0 D accommodation. Figure 3-21 tabulates the different types of abnormal color vision, their causes, and their prevalence. For most activities, abnormal color vision causes few problems, either because the exact identification of color is unnecessary or because there are other cues by which the necessary information can be obtained (e.g., relative location in traffic signals). Abnormal color vision does become a problem when color is the sole or dominant means used to identify objects, for example, in some forms of electrical wiring. People with abnormal color vision have difficulty with such activities.

Figure 3-19. Chromatic aberration: the optical power necessary to correct the focus of an eye for differences in refraction at different wavelengths (zero correction is arbitrarily set at 589 nm).

Figure 3-20. Chromatic aberration: because the eye's index of refraction is greater for short wavelengths than for long wavelengths, the eye focuses short wavelengths (F′B) in front of long wavelengths (F′R). Where self-luminous colors are used as signals, colored lights should be restricted to those that can be distinguished by people with the more common forms of color abnormality. The CIE has recently recommended areas on the CIE 1931 Chromaticity Diagram within which red, green, yellow, blue, and white signal lights should lie (see Figure 26-17 in Chapter 26). These areas are designed so that the red signal will be named as red and the green as green, even by dichromats, who are missing either a long or middle-wavelength photoreceptor pigment.31 It should be noted that for people with the most common form of abnormal color vision, the anomalous trichromats, the ability to discriminate colors shows wide individual differences. Some anomalous trichromats are barely distinguishable for people with normal color vision, whereas others resemble dichromats in their ability to discriminate colors.

Figure 3-21. The Classification, Characteristics, and Prevalence of Defective Color Vision Aging. As the visual system ages, a number of changes in its structure and capabilities occur.18 Usually, the first obvious change is loss of accommodation. Accommodative function decreases rapidly with age, so that by age 45 most people can no longer focus at near-working distances (approximately 40 cm) and might need optical assistance. This is known as presbyopia. By age 60, there is very little accommodative ability remaining in most of the population (Figure 3-22), resulting in a fixed-focus optical system. This lack of focusing ability is compensated somewhat by the physiologically smaller pupils in the elderly (senile myosis) because these increase the depth of field of the eye. However, the smaller pupils in turn increase the requirement for task luminance to maintain the same retinal illuminance as when the pupils were larger.

Figure 3.22. The decrease in the amplitude of accommodation with age. While the increasing rigidity of the lens, as well as many other forms of focusing difficulty, can be compensated by adjusting the optical power of the eye's optical system with spectacles and contact lenses, the other changes that occur in the eye cannot. As the visual system ages, the amount of light reaching the retina is reduced, more of the light entering the eye is scattered, and the spectrum of the light reaching the retina is altered by preferential absorption of the short visible wavelengths. The rate at which these changes occur accelerates after age 60. In addition to these changes in the optical characteristics of the eye, deterioration in the neurological components of the visual system also occurs in later life.18 The consequences of these changes with age are reduced visual acuity, reduced contrast sensitivity, reduced

color discrimination, increased time taken to adapt to large and sudden changes in luminance, and increased sensitivity to glare.18,32,33 Lighting can be used to partially compensate for these changes. Specifically, in an extensive long-term field study, the quality of life of the elderly has been shown to be improved by increasing the quality of their lighting.34 This raises the question of how to improve the quality of lighting for the elderly. Simply providing more light might not be enough. The light must be provided in a way that both disability and discomfort glare are controlled and veiling reflections are avoided. Where elderly people are likely to be moving from a well-lighted area to a dark area, such as a supermarket to a parking lot, a transition zone with a gradually reducing illuminance is desirable. Such a transition zone allows their visual system more time to make the necessary changes in adaptation. Partial Sight. Partial sight is a state of vision that falls between normal vision and total blindness. While some people are born with partial sight, the majority of people with partial sight are elderly. Among the partially sighted, 20% became partially sighted between birth and 40 years, 21% between 41 and 60 years and 59% after 60 years of age.32 Surveys in the United States and the United Kingdom suggest that the proportion of the total population who are classified as partially sighted are in the range 0.5 to 1%.35,36 The three most common causes of partial sight are cataract, macular degeneration, and glaucoma.34 These causes involve different parts of the eye and have different implications for how lighting might be used to help people with partial sight. 1. Cataract. This is an opacity developing in the lens. The effect of cataract is to absorb and scatter more of the light passing through the lens. This increased absorption and scattering occurring in the lens results in reduced visual acuity and reduced contrast sensitivity over the entire visual field because the scattered light degrades the contrast of the retinal image. This is known as disability glare, which occurs when light is scattered in the eye. The extent to which more light can help a person with cataract depends on the balance between absorption and scattering. More light will help overcome the increased absorption but if scattering is high, the consequent deterioration in the luminance contrast of the retinal image will reduce visual capabilities. The use of dark backgrounds against which objects are to be seen will also help.37,38 2. Macular degeneration. This occurs when the macular photoreceptors and neurons become inoperative due to bleeding or atrophy. The fovea is at the center of the macula lutea, and any loss of vision implies a serious reduction in visual acuity, color vision, and contrast sensitivity at high spatial frequencies. Typically, these changes make reading difficult, if not impossible. However, peripheral vision is largely unaffected so wayfinding is unchanged. Providing more light, usually by way of a task light, will help people in the early stage of macular degeneration to read, although as the deterioration progresses, additional light is less effective. Increasing the visual size of the retinal image by magnification or by getting closer is helpful at all stages, because this can increase the size of the retinal image sufficiently to reach parts of the retina beyond the macula. 3. Glaucoma. Glaucoma is due to an increase in intraocular pressure that damages the retina and the anterior optic nerve. Glaucoma is shown by a progressive narrowing of the visual field, which continues until complete blindness occurs or the intraocular pressure is reduced. As glaucoma develops, in addition to a reduction in visual field size, poor night vision, slowed transient adaptation, and increased sensitivity to glare occur, all due to the destruction of peripheral photoreceptors and neurons. However, the resolution of detail seen on axis is unaffected until the final stage. Lighting has limited value in helping people in the early stages of glaucoma, because where damage has occurred, the retina has been destroyed. However, consideration should be given to providing enough light for exterior lighting at night to enable the fovea to operate. Such lighting will be helpful only if glare is controlled. While the benefits of additional light depend on the specific cause of partial sight, there is one approach that

is generally useful for all those with partial sight. This approach is to simplify the visual environment and to make its salient details more visible. Details can be made more visible by increasing their size, luminance contrast, and color difference. As an example, consider the problem of how to set a table so that a person with partial sight can eat with confidence. The plate holding the food and the associated cutlery can be made more visible by using a contrasting tablecloth, e.g., a dark tablecloth with a white plate and cutlery. The food on the plate can be made easier to identify by using an overlarge plate so that individual food items can be separated from each other. The whole scene can be simplified by using solid colors rather than patterns. This same approach of simplification and enhanced visibility of salient information can be applied to whole rooms, for example, by painting a door frame in a contrasting color to the door so that the door is easily identified. Advice on designing lighting for the partially sighted is given in CIE Technical Report 123.37

Figure 3-23. A frequency of seeing function. As luminance contrast (see Equation 3-6) is increased, the number of times it is correctly seen relative to the number of times it is presented increases to a maximum of 100%.

THRESHOLD VISUAL PERFORMANCE Measurements of threshold visual performance are concerned with the limits of the visual system's capabilities. As an example of a threshold, consider the measurement of the minimum difference in luminance that can be detected between a region and an otherwise uniform surround. This function has been studied in great detail39 and has been used to relate the probability of detecting a small disc test object on a uniform background to the luminance contrast of the disc and the luminance of the background (Figure 323). As the luminance contrast of the disc is raised, the probability of seeing increases until, at a certain contrast, it can be detected 100% of the time. The luminance contrast at which the object can be detected 50% of the time is conventionally called the threshold luminance contrast. Threshold visual performance measurements can be made for the ability to resolve detail, to detect luminance differences and color differences, and to see temporal changes in luminance. All such threshold measurements depend greatly on the characteristics of the lighting, the task, and the visual system of the observer. Among the variables that have been shown to be important are:

Retinal illumination to which the visual system is adapted Spectral content of the illuminant Light distribution around the target Visual size of target (in units of angle or solid angle) Visual size of background (in units of angle or solid angle) Luminance of the target Luminance of the immediate background Luminance contrast of the target Color of the target Color of the background

Color difference between the target and the background Duration of exposure Temporal frequency characteristics Location of the target relative to the line of sight Movement of the target in the field of view Retinal image quality, as determined by the state of accommodation, pupil size, light scatter, and lens fluorescence

Additionally, such cognitive factors as attention, expectation, and habituation affect the measurement of threshold detectability and recognition of targets. The practitioner can control such variables as the illuminance, the light spectrum, and the light distribution. These variables are sometimes important, per se, and can affect such other important factors as target luminance contrast and color contrast. Occasionally, the practitioner can influence such task variables as target size and duration of exposure. Extensive details of the threshold performance of the visual system are given in Reference 40.

Some Definitions Retinal Illuminance. For a given individual looking at a given scene, the illuminance on the scene governs the luminances of the surfaces and hence the retinal illuminance. Luminances in object space can be related to retinal illuminance by the following function:

where Er = retinal illuminance in lm/m2, τ = ocular transmittance, θ = angular displacement of surface from the line of sight, k = constant whose value is 15, et = amount of light entering the eye in trolands, and is calculated by

where L = surface luminance in cd/m2, p = pupil area in mm2. It should be noted that the amount of light entering the eye, et, measured in trolands, is often referred to as retinal illumination. This is misleading because it does not take into account the transmittance of the ocular media and therefore does not represent the luminous flux density on the retina. Visual Size. For a target to be seen, it has to be larger than a minimum size. The relevant size of a target is an angular measure and depends on the physical dimensions, d, of the object itself; the angle of inclination, θ, of the target from normal to the line of sight; and the distance from the viewer, l. Size can be measured in a plane of two dimensions as a visual angle or in a volume in three dimensions as a solid angle, as shown in Figures 3-24a and 3-24b, respectively. Visual angle of an object can be approximated by the following equation:

For small angles, this can be simplified to:

Solid angle is given by the following equation:

The choice between visual angle and solid angle as a measure of visual size is largely determined by the symmetry of the object. Where the object is radially symmetrical such as a disc, or symmetrical in one dimension such as a grating, then visual angle is all that is required. Where there is no symmetry in the object, such as a letter F, solid angle is a better measure of visual size. Many times, visual angle or solid angle will be small. In this situation, visual size is measured in minutes of arc (60 minutes equals 1 degree) and solid angle is measured in microsteradians.

Figure 3-24. Dimensions required for calculating (a) visual angle and (b) solid angle. Luminance and Luminance Contrast. Given that a target is above the minimum size, it will be visible only if it differs from its immediate background in luminance or color. If it differs in luminance from the immediate background, the target has a luminance contrast. Luminance contrast is defined in several ways:

where

Lt = luminance of the target, Lb = luminance of the background. This equation results in luminance contrasts that range between 0 and 1 for targets that are darker than their backgrounds, and between 0 and infinity for targets that are brighter than their backgrounds. This equation is used most often in the former case, where the background is brighter than the target (e.g., printed text).

where Lg = greater luminance, Ll = lesser luminance. This equation results in contrasts between 0 and 1 for all objects, whether brighter or darker than their backgrounds. It is especially applicable in a situation like a bipartite pattern in which neither of the areas on the two sides of the border can be identified as target or background.

where L max = maximum luminance, L min = minimum luminance. The quantity defined by this equation is often called contrast, or Michelson contrast, but is usually and more properly called modulation. It gives a value between 0 and 1 for all objects. It applies to periodic patterns, such as gratings, which have one maximum and one minimum in each cycle. Because there are several different definitions of luminance contrast and different definitions have different ranges of possible values, it is important to know which definition is being used when the contrast of a target is specified. When a target and its background are both diffuse reflectors, the luminance contrast is not affected by changing the illuminance, so the luminance contrast can be calculated from the reflectances. However, if either the object or the background are directional reflectors, luminance must be used to calculate contrast. It should be noted that for calculating luminance contrast, it does not matter how the luminance is achieved. It makes no difference whether the luminance is produced by reflection from a surface, such as print; from a self-luminous source, such as a VDT screen; or by some combination, such as a display on a VDT screen with a reflected image superimposed. Color Difference. Visual targets that are larger than the minimum size but have the same luminance as the immediate background, that is, zero luminance contrast, can still be discerned by differences in color. Color difference can be calculated as a distance between the colors of the object and the immediate background using either the CIELAB or the CIELUV color spaces described in Chapter 4, Color, or other approximately uniform color space. It should be noted that one dimension of these color spaces is luminance, so a difference expressed across these color spaces includes both luminance and color difference. To measure only color differences, the separation of the color of the object and its immediate background on a plane of constant lightness in the CIELAB or CIELUV color spaces or the two-

dimensional CIE 1976 u',v' diagram can be used.

Spatial Resolution Spatial Summation. In complete darkness, the smallest amount of light that can be detected varies inversely with the area over which the light occurs. In other words, the total number of photons received in unit time, for detection, is constant. This relationship, which is known as Ricco's Law, takes the form:

where I = threshold luminous flux, measured in photons per unit area per unit time, A = target area. Spatial summation is relevant only for very small targets. Above a certain size, spatial summation ceases so that further increases in size make no difference to the smallest amount of light that can be detected. This critical size varies with location on the retina. For foveal vision, it is approximately 6 min of arc, increasing to approximately 0.5° for a target 5° off axis, and 2.0° for a target 35° off axis. Visual Acuity. The word "acuity" is used to describe the ability to resolve fine details. Several different kinds of acuity are recognized. 1. Resolution acuity. The ability to detect that there are two stimuli, rather than one, in the visual field. It is measured in terms of the smallest angular separation between two stimuli that can still be seen as separate, such as two nighttime stars. Typically, resolution acuity is of the order of 1 min of arc. 2. Recognition acuity. The ability to correctly identify a visual target, as in differentiating between a G and a C. Visual acuity testing performed using letters, as is done clinically, is a form of recognition acuity testing. Typically, recognition acuity is of the order of minutes of arc. 3. Vernier acuity. The ability to identify a misalignment between two lines. Vernier acuity is typically of the order of seconds of arc. Several examples of acuity test objects are shown in Figure 3-25. Gratings, letters, and Landolt rings have all been used as acuity test objects. As with many other threshold tasks, visual acuity varies with retinal illuminance, size of background field, exposure duration and target motion. It also varies with luminance contrast, but by convention acuity is measured only at high luminance contrast. In general, acuity is finest when the target falls on the fovea (Figure 3-26) and improves as the retinal illuminance increases, because increasing the retinal illuminance decreases receptive field size. As for the size of the background field, Lythgoe41 has shown that acuity continues to improve with background luminance as long as the background is large; when the background field is small, there is an optimum luminance for visual acuity, above which acuity declines (Figure 3-27). Visual acuity also increases as the exposure duration increases, up to approximately 500 ms, after which no further improvement occurs (Figure 3-28). Target movement can limit the exposure duration and the ability to keep the retinal image on the fovea. As might be expected, increasing target speed tends to reduce visual acuity (Figure 3-29). The only condition under which the fovea fails to have the best visual acuity is scotopic vision. In this condition, the fovea is inactive and the best visual acuity is found a few degrees off the line of sight.

Figure 3-25. Commonly used test objects for determining resolution limits and visual acuity. The critical size is represented by the dimension d.

Figure 3-26. Minimum resolution in minutes of arc, as a function of angular separation from the fovea. Three different targets were used: Landolt rings at a background luminance of 2.45 cd/m2 (open circles); Landolt rings at a background luminance of 245 cd/m2 (filled circles); sinewave gratings at a background luminance of 1118 cd/m2 (filled squares).

Figure 3-27. Effect of background luminance on visual acuity. The targets are Landolt rings on a background field measuring 0.85° by 1.7°. When the luminance of the surround field (S) equals the luminance of the target background (B) visual acuity continues to improve as background luminance increases.

Figure 3-28. Minimum spatial resolution in minutes of arc plotted against target exposure duration. Resolution improves as exposure duration increases up to about 500 ms. Longer exposure durations do not affect minimum resolution.

Figure 3-29. Minimum spatial resolution in minutes of arc plotted against angular velocity of target and observer movement.

Contrast Threshold The visual system gives virtually no useful information when the retina is uniformly illuminated, but is highly specialized to gather information about luminous edges and gradients in the visual field.

The ability to detect a target against a background can be quantified by its threshold contrast. By convention, threshold contrast is the luminance contrast of the target that can be detected on 50% of the occasions it is presented (Figure 3-23). Many factors affect threshold contrast. Among the more important are target size and retinal illuminance. Figure 3-30 shows the change in contrast threshold for a 4 min arc disc displayed for 200 ms plotted against adaptation luminance, for people of two different age groups. It shows that as adaptation luminance increases, the contrast threshold decreases, rapidly at first and then more slowly.33,42 Targets of different sizes exposed for different times give different absolute values of contrast threshold but all follow the same trend.

Contrast Sensitivity Function Visual acuity and threshold contrast separately define two aspects of a target that defines its visibility. Visual acuity sets the minimum size for a target to be seen and threshold contrast sets the minimum luminance contrast that is required for a target of a given size to be seen. The contrast sensitivity function combines these two measures by showing the minimum contrast required for targets of different sizes to be seen. Specifically, the contrast sensitivity function is a plot of contrast sensitivity against spatial frequency (Figure 3-31). It is usually based on data collected from grating targets of different spatial frequency. Spatial frequency is the reciprocal of the visual angle of one period of the grating and is measured in cycles / degree. Contrast sensitivity for a given spatial frequency is the reciprocal of the luminance contrast of the grating at threshold. Targets that have a spatial frequency and contrast sensitivity such that they lie above the contrast sensitivity function are invisible (i.e., can be detected on fewer than 50% of the occasions presented) and those that lie below the contrast sensitivity function are visible (i.e., can be detected on more than 50% of occasions presented). For complex targets that contain many different spatial frequencies, the contrast sensitivity function can be used to determine if and how the target will appear by breaking it into its spatial frequency components.43 The target will be visible only if at least one spatial frequency component has a contrast sensitivity less than the contrast sensitivity function. Exactly how the target will appear will depend on the weighting given to each of its spatial frequency components by the contrast sensitivity function.

Figure 3-30. Threshold contrast data for a group of 60- to 70-year-olds (x) compared to the threshold contrast curve for a group of 20- to 30-year-olds (solid line), as a function of luminance. The dashed curve is the same as the solid curve but displaced upward by a factor of 2.51. Threshold contrast was calculated according to Equation 3-6.

Figure 3-31. The spatial contrast sensitivity function for foveal vision, at different target luminances. Many factors affect the contrast sensitivity function. Among the most important are the adaptation luminance, the location in the visual field, and the number of cycles of the stimulus. Figure 3-31 shows the variation of contrast sensitivity function with adaptation luminance. As the adaptation luminance changes the operating state of the visual system from scotopic to photopic, the contrast sensitivity increases for all spatial frequencies; the spatial frequency at which the peak contrast sensitivity occurs increases, and the highest spatial frequency that can be detected increases. As for the effect of the location in the visual field, contrast sensitivity is reduced at all spatial frequencies with increasing eccentricity, but the decrement is greater for high spatial frequencies. More details of the changes in contrast sensitivity functions with different lighting and visual conditions and examples of its diagnostic use can be found in Reference 40.

Temporal Resolution Just as the visual system responds to variations of luminance in space, it also responds to variations of luminance in time. Temporal Summation. For single brief flashes of light (less than 100 ms), any combination of luminance (L) and flash duration (t) with the same product produces the same perception. This characteristic is known as Bloch's law:

For single brief flashes of light longer than approximately 100 to 200 ms, the perception of the flash is solely a function of stimulus luminance (Figure 3-32). Flicker. As a repetitive flashing stimulus is increased in frequency, it eventually reaches a point where it is perceived as steady rather than as intermittent; this is the critical flicker frequency (or critical fusion frequency, CFF). The frequency at which the fusion occurs varies with stimulus size, shape, retinal location, adaptation luminance, and modulation depth. Figure 3-33 shows the relationship of CFF to adaptation luminance for centrally fixated test objects of different sizes. The CFF rarely exceeds 60 Hz even for a large visual area with 100% modulation, seen at a high adaptation luminance. This is just as well because all light sources that operate from an ac electrical supply show some fluctuation in light output. The electrical

supply frequency in North America is 60 Hz, which means the fundamental frequency of oscillation in light output is 120 Hz although there might be some 60 Hz component present. The fundamental frequency of the light output oscillation (120 Hz) is twice the fundamental frequency of the electrical supply (60 Hz) because of the positive and negative halves of the ac cycle.

Figure 3-32. Total number of quanta for seeing a flash of light as a function of the duration of the flash.

Figure 3-33. Critical flicker frequency (CFF) as a function of source area and retinal illuminance. Another way of considering the combined effects of temporal modulation and temporal frequency is through the temporal modulation transfer function (MTF). This is the equivalent in time of the spatial contrast sensitivity function (Figure 3-31). Figure 3-34 shows the temporal MTF for different adaptation luminances. The vertical axis is the reciprocal of percent temporal modulation and the horizontal axis is the frequency of fluctuation measured in cycles per second. Figure 3-34 shows that in photopic conditions (i.e., above approximately 3 cd/m2), the visual system is most sensitive to frequencies in the range 10 to 30 Hz and that as the adaptation luminance decreases, the absolute sensitivity to flicker decreases, the frequency at which the peak sensitivity occurs decreases, and the highest frequency that can be detected decreases. These temporal modulation transfer functions, and others for different conditions, can be used to determine the likelihood that a given fluctuation in light will be perceived as flickering. For a fluctuation with a complex waveform to be seen as flicker, at least one of its frequency components must have a modulation sufficiently high that the modulation sensitivity is below the temporal MTF. Knowledge of the visual system's temporal response is most helpful when considering the detection of flashing signals and the perception of animated signs.

Sensitivity to flicker differs across the retina. The fovea can follow flicker rates up to approximately 60 Hz at moderate luminances, but is relatively insensitive to low frequency modulations. The peripheral retina, on the other hand, can detect flicker rates to approximately 15 Hz, but is very sensitive to small flicker amplitudes. This is why flicker is often detected in the peripheral field but disappears when the light is viewed directly.

Figure 3-34. Temporal modulation transfer (contrast sensitivity) function for different adaptation luminances for a 68° field of view. It is widely recognized that visible fluctuations in light occurring over a large area can cause visual discomfort or annoyance. Recently, however, reductions in the prevalence of headaches have been reported when fluorescent lamps were operated on high-frequency electronic ballasts as compared to when they were operated on conventional (British) 50-Hz magnetic ballasts, even though flicker was not visible. This implies that flicker might have subliminal effects on the visual system. This hypothesis is supported by electrophysical recordings under such conditions.44

Color Discrimination The visual system varies in its ability to discriminate among wavelengths. There are regions of maximum wavelength discrimination in the middle of the visible spectrum but discrimination falls off rapidly at the spectral extremes.45 Likewise, the ability to discriminate hue from white is wavelength dependent. Monochromatic colors from the ends of the visual spectrum are more easily discriminated from white because they are more saturated than colors in the middle of the spectrum.46 The ability to discriminate nonspectral colors is also related to their chromaticities.47 Generally, color discrimination is best in the fovea and decreases toward the periphery. However, color discrimination for very small fields (20 min of arc or less) presented to the fovea is poor because there are very few short-wavelength S-cones in the center of the fovea. This effect is known as small-field tritanopia.47

The ability to discriminate between colors can be estimated in terms of distances in a uniform 3-D chromaticity space (see Chapter 4, Color). MacAdam48 produced a series of ellipses around the chromaticity coordinates of a number of different colors (Figure 3-35). Each ellipse sets the boundary at which a given percentage of people are able to determine that two colors, one with chromaticity coordinates at the center of the ellipse and one with chromaticity coordinates on the ellipse, are just noticeably different. Full details are given in Reference 8. MacAdam's ellipses were determined in conditions that offer the maximum sensitivity to color differences: side-by-side comparison, unlimited observation time, foveal viewing, and photopic operation of the visual system. Changing any of these factors and adding distracting or confusing stimuli can be expected to increase the difference in color needed to reach discrimination threshold. Of particular importance is the amount and spectral power distribution of the light reaching the retina. The retinal illuminance is important because it determines the operating state of the visual system. If the retinal illuminance is in the scotopic range, no colors can be seen and no discrimination is possible. In the mesopic range, colors can be seen but the discrimination of colors is poor, particularly for low reflectance colors. The color discrimination threshold is reduced as the retinal illuminance increases. This process continues until a retinal illuminance of approximately 30 trolands is reached, after which there is little change in the ability to discriminate colors. The light spectrum is important because it changes the stimulus to the visual system.

SUPRATHRESHOLD VISUAL PERFORMANCE Threshold visual performance deals with what can just be seen. Suprathreshold visual performance is concerned with tasks that are visible because their important aspects are well above threshold levels. This raises the question as to why lighting conditions make a difference to task performance once what has to be seen is visible. The answer is that although the stimuli are visible, lighting influences the speed and accuracy with which the visual information extracted from the stimuli can be processed. Like threshold visual performance, suprathreshold visual performance is governed by such parameters as retinal illuminance, task contrast, visual size, and the characteristics of the visual system. Retinal illuminance is largely determined by the luminance of the visual field that is viewed and hence by the illuminance on the surfaces that form that field.

Figure 3-35. The 1931 CIE chromaticity diagram showing a selection of MacAdam ellipses, enlarged by a factor of ten. From G. Salvendy, Handbook of human factors and ergonomics, 2nd ed. Copyright © 1997. Reprinted by permission of John Wiley & Sons, Inc. One approach to studying suprathreshold visual performance is to examine task performance for a variety of realistic tasks requiring vision. Several studies have been conducted mimicking realistic tasks to determine how illumination affects performance.49-53 This approach allows the experimenter to assess performance for a specific task in suprathreshold conditions, but it is difficult to generalize the results with high precision to other, even superficially similar tasks because it is impossible to separate visual from nonvisual components of performance. An example of nonvisual components would be the time taken to turn the page in a proofreading task (see "Visual Performance, Task Performance, and Productivity" below for a discussion of task structure). Another approach was developed by Blackwell based on his extensive psychophysical measurements of threshold contrast.33,39,42 He developed several models for predicting suprathreshold performance from threshold performance.54-56 These systems all used the same concept as the basis of prediction, namely a simple multiplier derived from the ratio of the actual contrast presented by a target to the threshold contrast at the same adaptation luminance. This multiplier is known as Visibility Level. Unfortunately for simplicity, it was later shown that it is not possible to predict suprathreshold performance accurately from threshold performance, because if the stimuli at threshold are different, the same Visibility Level produces different suprathreshold performances.57 Further, the complexities introduced by nonvisual components and off-axis working were not fully appreciated. As a result, the number of factors that had to be introduced to make the predictions fit a set of experimental data increased dramatically as the number of data sets increased, eventually leading to a loss of credibility. As a result, attempts to use the systems have been few. As early as 1935, Weston58,59 recognized the importance of a systematic, direct study of suprathreshold

performance utilizing variables that had been shown to be important to threshold vision, namely target size and target luminance contrast seen at different background luminances. The curves in Figure 3-36 demonstrate the effects of illuminance on detection of Landolt rings of different orientations and printed in different contrasts and sizes.58-59 Performance was defined, in these studies, as an aggregate score based on speed and accuracy.

Figure 3-36. Mean performance scores for Weston's Landolt ring charts of different visual size and luminance contrast (see Equation 3-6), plotted against illuminance. The performance score for each individual is given by the expression

An analysis of Weston's work60 identified several design and analysis flaws, including a lack of documentation about specific visual characteristics and the use of a scoring system that failed to include correct rejections as part of the accuracy metric. Weston's performance data shown in Figure 3-36 can provide only general trends in suprathreshold response but, importantly, trends that cannot be gleaned from a knowledge of threshold vision. In general, Weston showed that as background luminance increased, performance (measured in terms of speed and accuracy) increased rapidly at first but then less and less until a point was reached where very large changes in background luminance were required to make very small changes in performance. This trend of diminishing returns was more pronounced for high-contrast, large targets than for low-contrast, small targets. He also showed that performance for a small, low-contrast target could not be brought to the same level as a large, high-contrast target simply by increasing illuminance. Rather, changing the size and luminance contrast of the target often had a much larger effect on suprathreshold visual performance than increasing the illuminance over any practical range. Several studies of suprathreshold performance have extended Weston's approach.50-53,61 All of these studies have produced results consistent with the general trends shown by Weston.

Models of On-Axis Visual Performance Weston's results illustrate the general form of the relationship between the visual size, luminance contrast of the target, and the retinal illuminance. Other researchers have provided quantitative models using more precise techniques. The general trends of suprathreshold performance, shown by Weston, have not been

contradicted by these models.

Figure 3-37. Relative visual performance (RVP) plotted as a function of task contrast (see Equation 3-7) and retinal illuminance (in trolands) for several different target sizes measured as solid angle (microsteradians).

Figure 3-38. Reading speed in words/second plotted against visibility level (size).

The Relative Visual Performance (RVP) model of visual performance is a quantitative model based on an extensive dataset made up of the changes that occur in reaction time for the detection of visual stimuli seen by the fovea.62-65 The conditions covered in the dataset represent a wide range of adaptation luminances, luminance contrasts, and visual sizes. By using simple reaction time as a measure, this model attempts to minimize the nonvisual components in the task. By basing the model on the difference in reaction time that occurs for different combinations of adaptation luminance, luminance contrast, and visual size, the effect of any remaining nonvisual components is further minimized. Therefore, the RVP model shows the effect of adaptation luminance, luminance contrast and visual size on suprathreshold visual performance undiluted by nonvisual components. Figure 3-37 shows the form of the relative visual performance (RVP) model for four different visual size tasks, each surface being for a range of luminance contrasts and retinal illuminances. The overall shape of the relative visual performance surface has been described as a plateau and an escarpment.63 In essence, it shows that the visual system is capable of a high level of visual performance over a wide range of visual sizes, luminance contrasts, and retinal illuminations (the plateau) but at some point either visual size, luminance contrast, or retinal illumination become insufficient and visual performance collapses rapidly (the escarpment) towards a threshold state. The RVP model provides a quantitative means of predicting the effects of changing either task size, luminance contrast, or adaptation luminance for on-axis, suprathreshold visual performance. It is applicable to luminances in the photopic range but does not take into consideration the effect of reduced retinal image quality caused by limited accommodation, nor the effect of color differences between the target and the background. It can be only applied once a decision is made as to what constitutes the true critical size of the target. The RVP model has been validated in that it has been shown to predict the form of the change in performance produced by different lighting conditions, measured in three independent experiments, using different visual tasks.60,66,67 It can be applied using input variables that can all be measured directly from the task. The RVP model is based on reaction time data for detecting the presence of a square target. Such a target requires contrast discrimination but does not require resolution of detail. A task that does require resolution of detail is reading. Figure 3-38 shows task performance measured as reading speed plotted against the Visibility Level of the letters being read.66 In this case, Visibility Level was defined as the ratio of the actual size to the threshold size of the letters of a given luminance contrast and at a set adaptation luminance. Figure 3-38 shows that reading speed changed little until the visibility level fell below a value of 3 but declined rapidly as the resolution threshold (VL = 1) was approached. Here too, there was a plateau and escarpment of task performance. For these reading speed data, a function based on Visibility Level was found to fit the data slightly better than the RVP model, most of the difference occurring for print sizes smaller than 6 point, where resolution might be expected to be important. Despite theoretical arguments,68 in many ways the results of these studies of suprathreshold visual performance and task performance are more remarkable for their similarities than their differences. All show a plateau and escarpment form. This has important implications. The existence of a plateau of visual performance implies that for a wide range of visual conditions, visual performance changes little with changes in the lighting conditions. However, the RVP model does show that the plateau is not completely flat, there is a slight improvement in visual performance as adaptation luminance is increased, even when luminance contrast is high and visual size is large. What this means is that for many visual tasks, suprathreshold visual performance is insensitive to lighting conditions, but if it is required to maximize visual performance, increasing adaptation luminance can be effective.

Figure 3-39. The probability of detection for targets of (a) contrast 0.058, size = 19 min. arc; (b) contrast = 0.08, size = 10 min. arc; (c) contrast = 0.044, size = 10 min. arc; within a single fixation pause, plotted against deviation from the visual axis. Each curve can be used to form a "visibility lobe" for each target by assuming symmetry about the visual axis. It should also be noted that the RVP model is based on the luminance contrast presented to the observer, regardless of how that contrast is achieved. This means that both light polarization and distribution can affect visual performance for tasks that involve specularly reflecting materials, because both can change luminance contrast.60,69 Light distribution can produce veiling reflections (see "Lighting Conditions That Can Cause Discomfort" below) that can make luminance contrast larger or smaller, depending on the specific arrangement of the materials. The change in luminance contrast can be large but it is difficult to control because it depends critically on the geometry between the source of luminance being reflected, the task, and the observer. A small change in position of any of these entities can markedly change the luminance contrast.61 As for light polarization, in principle, polarized light is capable of eliminating specularly reflected light, but this too is very dependent on the geometry between the source of polarized light, the reflecting surface and the observer, as well as the magnitude and nature of the polarization.69 A discussion on the physics of polarization is provided in Chapter 1, Light and Optics.

Visual Search The RVP model discussed above is applicable to tasks that are imaged on the fovea. This is likely to be the situation when the observer knows where to look (e.g., reading). However, there is a class of tasks in which the object to be detected (i.e., target) can appear anywhere in the visual field (e.g., driving or industrial inspection). These tasks involve visual search. Visual search is typically undertaken through a series of eye fixations, the fixation pattern being guided either by expectations about where the target is most likely to appear or by what part of the visual scene is most important. Typically, the target is first detected in the periphery of the retina. Detection is followed by eye movements that bring the detected target onto the fovea, where it is identified. The speed with which a visual search task is completed depends on the size, luminance contrast, and color difference of the target; the presence of other targets in the search area; and the extent to which the target is different from the other targets. The simplest visual search task is one in which the target appears somewhere in an otherwise empty field, such as paint scratches on a car body. The most difficult visual search task is one in which the target is situated in a cluttered field, where the clutter is very similar to the target to be found, such as searching for a particular face in a crowd. The speed of visual search is determined by both the task characteristics and the lighting conditions. The task characteristics that hasten visual search are those that make the target stand out from its background (i.e., make it visible) and make it different from surrounding clutter (i.e., make it conspicuous). To make a target visible, its visual size and luminance contrast must be well above the threshold values. To make a target conspicuous, it should differ from the surrounding clutter on as many perceptual dimensions as possible. Among such dimensions are size, shape, color, movement, and flicker.

The extent to which a lighting installation is effective in revealing a target can be estimated from the object's visibility lobe.70 The visibility lobe is the distribution of the probability of detecting the object within one fixation pause (Figure 3-39). This probability is at maximum when the target is viewed with the fovea and decreases with increasing eccentricity from the fovea. The probability distribution is assumed to be radially symmetrical about the visual axis, resulting in circular contours of equal probability of detection within one fixation pause around the fixation point. Given that the interfixation distance is related to the visibility lobe and the search area is fixed, the time taken to find a target is inversely related to the size of the visibility lobe. For objects that appear on a uniform field, the visibility lobe is based on the detection of the object. For objects that appear among other similar objects, the visibility lobe is based on the discriminability of the object from the others surrounding it. Visual search is fastest for targets that have the largest visibility lobe.

Effect of Spectral Content on Suprathreshold Performance The models of on-axis suprathreshold visual performance and visual search respectively discussed in "Models of On-Axis Visual Performance" and "Visual Search" ignore the possibility of the spectral content of the illuminant affecting the visual performance. This is reasonable given that previous research has shown little effect of spectral content on the performance of achromatic tasks.71 However, there is little doubt that the spectral content of the illuminant can affect the performance of tasks requiring color discrimination, nor that spectral content is important for visual search where color is one of the dimensions on which the target differs from other objects around it.72 In these situations, the spectral content of the illuminant changes the stimuli presented to the visual system. As a general rule, the higher the CIE General Color Rendering Index of the illuminant, the greater the color differences in the task and the easier it is to make the required discriminations. While the spectral content of an illuminant might be expected to be important for any task where the spectral content changes an important aspect of the stimuli the task presents to the visual system, there is little evidence that it is important for all tasks. This has not stopped claims being made for what are called full-spectrum lamps. These lamps, which have no widely accepted definition, are typically fluorescent lamps with spectral emission in all parts of the visible spectrum and in the near UV, with a correlated color temperature of 5000 K or more and a CIE General Color Rendering Index of 90 or more. Claims have been made that the use of such lamps benefit task performance, human health, and happiness. These claims have little merit in most cases.73 However, in recent years a number of studies by Berman and his colleagues have shown that spectral content can influence performance for achromatic resolution tasks such as reading or rapidly identifying gaps in a Landolt C acuity target (Figure 3-40).74-76 The proposed explanation of these findings rests on the role of pupil size. Specifically, pupil size in a large visual field is determined by the response of the rod photoreceptors, even in photopic conditions; the larger the response from the rods, the smaller the pupil area.77 The response from the rod photoreceptors can be increased by increasing the amount of short wavelength energy received at the eye. A smaller pupil area has three effects on the retinal image: it reduces the retinal illumination, it increases the depth of field, and it reduces aberrations. The first of these effects, the reduction in retinal illuminance, can be expected to degrade visual performance. The other two, increasing the depth of field and reducing aberrations, can be expected to improve the quality of the retinal image and hence to improve visual performance. All these effects are small, and how they trade off will depend on the inherent quality of the individual's optical system. An individual who is perfectly refracted will gain little from increasing the depth of field, so this person might be expected to experience deterioration of visual performance under a light source that produces smaller pupil sizes. However, most people do not have perfect refraction. For these people, the evidence suggests that light sources that do promote smaller pupil sizes can increase visual performance where the task conditions place it close to threshold, e.g., low luminance contrast, limited exposure time, and when the surfaces being viewed reflect most of the incident short wavelength light. Figure 5-18 gives spectral reflectances of a number of common building surfaces.

Visual Performance, Task Performance, and Productivity

Although our understanding of the effects of lighting conditions on visual performance has grown in recent years, it is important to realize that there is an inherent limitation on the generalizability of this understanding to the performance of all visual tasks. Figure 3-41 shows a conceptual relationship between visual stimuli, visual performance, task performance, and productivity. The stimuli to the visual system are determined by the task characteristics and the way the task is lighted. These stimuli and the operating state of the visual system determine visual performance. Most visual tasks have three components: visual, cognitive, and motor. The visual component refers to the process of extracting information relevant to the performance of the task using the sense of sight. The cognitive component is the process by which these sensory stimuli are interpreted and the appropriate action determined. The motor component is the process by which the stimuli are manipulated to extract information and the consequential actions carried out. As an example, consider the task of rectifying a software bug using an instruction manual. The visual component is to see the marks on the page of the instruction manual. The cognitive task is understanding what it means. The motor component is striking the right keys on the keyboard. Every task is unique in its balance between visual, cognitive, and motor components and hence in the effect lighting conditions have on task performance. It is this uniqueness that makes it impossible to generalize from the effect of lighting on the performance of one task to the effect of lighting on the performance of another. The RVP model of visual performance for on-axis tasks and the visual search model discussed above can be used to quantify the effects of lighting conditions on visual performance, but there is no general model to translate those results to task performance. Unfortunately, task performance is what is needed in order to measure productivity and to establish cost-benefit ratios comparing the costs of providing a lighting installation with the resulting benefits in terms of better task performance. When reading the literature, it is important to note that measures of task performance are sometimes erroneously called measures of visual performance.

Figure 3-40. Means and associated standard errors of the proportion of Landolt ring orientations, presented for 200 ms on a spectrally neutral background, which were correctly identified; plotted against luminance contrast (see Equation 3-7), for four different target background luminances: (a) 11.9, (b) 27.7, (c) 47.0, (d) 73.4 cd/m2. In all four diagrams, the upper curve is for a scotopically enriched illuminant (surround field scotopic luminance = 228 cd/m2), and the lower

curve is for a scotopically deficient illuminant (surround field scotopic luminance = 13 cd/m2). Both illuminants produce a surround field photopic luminance of 53 cd/m2.

Improving Visual Performance The main purpose of many lighting installations is to enable people to perform their work quickly, easily, comfortably, and safely. To achieve this aim, it is necessary to provide lighting that ensures people are operating on the plateau of visual performance and not on or close to the escarpment. Although the discussion above has been focused on lighting conditions, it is important to recognize that visual performance can be improved by changing the characteristics of the task as well as the lighting. The following list is divided into two parts: task changes and lighting changes. Not all the following suggestions apply in every situation, and not all are appropriate for every problem. Changing the Task

Increase the size of detail in the task, such as by magnification. Increase the luminance contrast of the detail in the task, for example, by adding toner to the printer. For off-axis tasks in a cluttered field, make the object to be detected clearly different from the surrounding objects on as many different dimensions as possible, such as by using size, contrast, color, and shape. Ensure the object presents a sharp image on the retina (for example, by going to an optometrist).

Changing the Lighting

Increase the adaptation luminance, such as by increasing the illuminance. Where good color discrimination is needed, select a lamp with a high CRI (e.g., greater than 80). Design the lighting so that it is free from disability glare and veiling reflections, such as by eliminating any direct views of the light source and by using matte materials.

Figure 3-41. A conceptual diagram of the complex relationships between the stimuli to the visual system and their impact on visual performance, task performance, and productivity. The arrows indicate the direction of their effects. The dotted arrow between visual performance and visual size indicates that if visual performance is poor, a common response is to move closer to the stimulus to increase its visual size. From G. Salvendy, Handbook of human factors and ergonomics, 2nd ed. Copyright © 1997. Reprinted by permission of John Wiley & Sons, Inc.

VISUAL COMFORT Lighting installations are rarely designed for visual performance alone. Visual comfort is almost always a consideration. The aspects of lighting that cause visual discomfort include those relevant to visual performance and extend beyond them.

Symptoms and Causes of Visual Discomfort Visual discomfort can give rise to an extensive list of symptoms. Among the more common are red, sore, itchy, and watering eyes; headaches and migraine attacks; gastrointestinal problems; and aches and pains associated with poor posture. Visual discomfort is not the only possible source of these symptoms. All can have other causes. This vagueness makes it essential to consider other possible causes before ascribing an occurrence of any of these symptoms to the lighting conditions. The visual system is designed to extract information from the visual environment. This is essentially a "signal-to-noise" problem with the signal being the information desired and the noise being all the other information in the visual environment. Features of the visual environment that reduce the signal-to-noise ratio, either by reducing the signal or increasing the noise, can cause visual discomfort. There are several different situations where signal-to-noise ratio problems can occur. Any visual task that has characteristics that place it close to threshold has a low signal-to-noise ratio and hence has a high level of visual difficulty. One reaction to a high level of visual difficulty is to bring the task closer to increase its visual size. As the task is brought closer, the accommodation mechanism of the eye adjusts to keep the retinal image in focus, an adjustment that might make it operate close to its limits. This adjustment can lead directly to fatigue of the eye muscles, and indirectly to fatigue of other muscles caused by the observer adopting an unusual posture. Such muscle fatigue can produce symptoms of visual discomfort. Even when it is not possible to move closer to the task, signals close to threshold can generate symptoms of visual discomfort. An example of this occurs when driving in fog or in a "whiteout" snowstorm. In both situations, the visual system is searching for information that is hidden but that may appear suddenly and require a rapid response. The stress while driving in these conditions is a common experience. Symptoms of visual discomfort may also occur when there is a high visual noise level. Looking at a printed page that has large areas of high-contrast gratings has been associated with the occurrence of headaches, migraines, and reading difficulties.78 Symptoms of visual discomfort also occur when there are several disassociated strong signals in the visual field. The visual system has a large peripheral field that detects the presence of targets, which are then examined using the small, high-resolution fovea. For this system to work, objects in the peripheral field that are bright, moving, or flickering have to be easily detected. If, upon examination, these bright, moving, or flickering objects prove to be of little interest, they become sources of distraction because their attention gathering power is not diminished after one examination. Ignoring objects that attract attention is stressful and can lead to symptoms of visual discomfort.79 Another form of this source of visual discomfort occurs when there are two alternative views of the world. This can happen when a lighting installation is reflected from a computer monitor. The monitor presents both the generated image and a reflected image of the room.

Lighting Conditions That Can Cause Discomfort There are many different aspects of lighting that can cause discomfort. Insufficient light for the easy performance of a task is an obvious problem that can be resolved by one of the approaches suggested in "Improving Visual Performance." Here, attention will be given to flicker, glare, shadows, and veiling reflections. It should be noted that the impact of these conditions on discomfort depends on the context. All can be used to positive effect in some contexts.

Flicker. A general lighting installation that produces visible flicker will be almost universally disliked, unless it is being used for entertainment. The magnitude of individual differences, and the fact that electrical signals associated with flicker can be detected in the retina even when there is no visible flicker,44 imply that a clear safety margin is necessary if flicker is not to be perceived by anyone. The main variables that determine flicker perception are the frequency and percentage modulation of the oscillation in light output, the proportion of the visual field over which the flicker occurs, and the adaptation luminance. Temporal modulation transfer functions (see "Temporal Resolution" above) can be used to predict whether a given fluctuation of light output will be visible. To eliminate the perception of flicker, it is necessary to increase the frequency of oscillation above the critical flicker frequency or to reduce the percentage modulation of the oscillation, the area of the visual field over which the oscillation occurs, or the adaptation luminance. The last two possibilities occur rarely with general lighting. A much more common approach is to use high-frequency control gear for discharge lamps and to mix the light from lamps powered from different phases of the electricity supply,80 both of which increase the frequency and reduce the modulation of oscillation in light output. The use of highfrequency control gear has been associated with a reduction in the prevalence of headaches.81 Although flicker occurring over a large area of the visual field is almost always disturbing, flicker occurring over a very small part of the visual field is much less disturbing and is a very effective way to draw attention to a particular location. This technique is widely used with emergency vehicles. Glare. Glare occurs in two ways. First, it is possible to have too much light. Too much light produces a simple photophobic response, in which the observer squints, blinks, or looks away. Too much light is common in full sunlight. The only solution to this problem is to reduce the retinal illuminance by obscuring a bright part of the visual field (e.g., by wearing a baseball cap) or by lowering the luminance of the whole visual field (e.g., by wearing sunglasses). Second, glare occurs when the range of luminance in a visual environment is too large. Glare of this sort can have two effects: a reduction in visual performance until it is close to or on the escarpment of visual performance (see "Models of On-Axis Visual Performance"), and a feeling of discomfort. Glare that reduces visual performance is called disability glare and is due to light scattered in the eye, reducing the luminance contrast of the retinal image. The effect of scattered light on the luminance contrast of the target can be mimicked by adding a uniform "veil" of luminance to the target. The magnitude of disability glare can be estimated by calculating this equivalent veiling luminance. Several investigators82-85 have examined the role of glare-source luminance and angular separation from the primary object of regard as producers of disability glare; they have each derived slightly different functions, but a commonly used expression is

where Lv = equivalent veiling luminance in cd/m2, Ei = illuminance from the ith glare source at the eye in lux, θi = angle between the target and the ith glare source in degrees. The effect of disability glare on the luminance contrast of the perceived target can be determined by adding the equivalent veiling luminance to all elements in the formulas for luminance contrast (Eq. 3-6). For example, to include the effect of disability glare on the luminance contrast of print, the formula is

where Lv = equivalent veiling luminance in cd/m2, Lt = target luminance in cd/m2, Lb = background luminance in cd/m2. Although disability glare is most commonly thought of as coming from discrete sources, such as oncoming automobile headlamps, every luminous point in space acts as a source of stray light for nearby points and reduces contrast, thereby making edges in the visual field less conspicuous. The illuminance at the eye term in Equation 3-11 integrates the scattering effects produced by stray light from all points. Disability glare is rarely important in interior applications but is common on roads at night from oncoming headlights and during the day from the sun. Disability glare usually also causes discomfort, but it is possible to have disability glare without discomfort when the glare source is large. This can be seen by looking at a picture hung on a wall adjacent to a window. The picture will usually be much easier to see when the eyes are shielded from the window. The other form of glare is called discomfort glare. Discomfort glare is a sensation of annoyance or pain caused by high luminances in the field of view. By definition, discomfort glare does not affect visual performance but does cause discomfort. While the cause of disability glare is well known (intraocular light scattering; see "Optical Components"), that of discomfort glare is not understood. Laboratory studies have related discomfort glare to pupillary and facial muscle activity,86-88 but the validity of these mechanisms as causes of discomfort glare is not yet widely accepted. Despite this lack of understanding, the desire for some method to determine if a lighting installation will produce discomfort glare has led to the development of a number of empirical prediction systems in different countries.89 In North America, the empirical prediction system is called the Visual Comfort Probability (VCP) system. This system is based on assessments of discomfort glare for different sizes, luminances, and numbers of glare sources, their locations in the field of view, and the background luminance against which they are seen, for conditions likely to occur in interior lighting. The criterion used to measure the effect of these variables is the luminance just necessary to cause discomfort, a threshold criterion termed the borderline of comfort and discomfort (BCD).90 The visual comfort probability (VCP) system evaluates lighting systems in terms of the percentage of the observer population that will accept the lighting system and its environment as being comfortable, using the perception of glare due to direct light from luminaires to the observer as a criterion. The following factors have been found to influence subjective judgments of discomfort glare:

Room size and shape Room surface reflectances Illuminances Luminaire characteristics Number and location of luminaires Luminance of the entire field of view Observer location and line of sight Differences in individual glare sensitivity

The VCP can be calculated for specific lighting systems and given observer lines of sight (see Chapter 9, Lighting Calculations). However, in order to systematize the calculations and to permit comparison of

luminaires, standard conditions have been adopted.91 These are:

An initial illuminance of 1000 lx (100 fc) Room surfaces with 80% for the effective ceiling cavity reflectance, 50% for the wall reflectance, and 20% for the effective floor cavity reflectance Mounting heights above the floor of 2.6, 3, 4, and 4.9 m (8.5, 10, 13, and 16 ft) A range of room proportions to include square, long-narrow, and short-wide rooms A standard layout involving luminaires uniformly distributed throughout the space An observation point 1.2 m (4 ft) in front of the center of the rear wall and 1.2 m (4 ft) above the floor A horizontal line of sight looking directly forward A limit to the field of view corresponding to an angle of 53° above and directly forward from the observer

Luminaire manufacturers use the VCP formulas and the standard conditions to produce tabular estimates of the level of discomfort glare produced by a regular array of their luminaires for a range of standard interiors. These tables provide all the precision necessary for estimating the level of discomfort glare likely to occur in interiors. By consensus, discomfort glare is not a problem in lighting installations if all three of the following conditions are satisfied:91

The VCP is 70 or more. The ratio of the maximum luminance (luminance of the brightest 6.5 cm2 [1 in.2] area) to the average luminaire luminance does not exceed 5:1 at 45°, 55°, 65°, 75°, and 85° from nadir for crosswise and lengthwise viewing. Maximum luminances of the luminaire crosswise and lengthwise do not exceed the following values.

The principal research used to establish the VCP system involved luminances of magnitude comparable to those produced by fluorescent lamps.90,92-98 Further, the most extensive field validation used lighting systems containing fluorescent luminaires. Although mathematically VCP can be applied to virtually any lamp and luminaire combination, extrapolation to lamps and luminaires with significantly different luminance patterns has not been validated. Therefore, the validity of applying VCP to clusters of small light sources, for example, is questionable. The VCP system is based on empirical relations derived from a variety of experiments. It has been concluded that differences of 5 units or less are not significant. In other words, if two lighting systems do not differ in VCP rating by more than 5 units, there is no basis for judging that there is a difference in visual comfort between the two systems. Artifacts introduced by using different computational procedures for two lighting systems can further spread the VCP values for two systems that are not reliably different.99 An alternative, simplified method of providing an acceptable degree of comfort has been derived from the formulas for discomfort glare. This method is based on the premise that luminaire designers do not design different luminaires for rooms of different sizes but rather consider the probable range of room sizes and design for the "commonly found more difficult" potential glare situation (in rooms less than 6 m [20 ft] in length and width, the luminaires are largely out of the field of view). This simplified method is only applicable to flat-bottom luminaires.100-102

While the VCP system is used in North America, the rest of the world uses somewhat different discomfort glare prediction systems. Nearly all these systems are based on a formula that implies that discomfort glare increases as the luminance and solid angle of the glare source at the eye increase and decreases as the luminance of the background and the deviation of the glare source from the line of sight increases.89 Comparative evaluations between the different discomfort glare prediction systems for a common range of installations have shown that their predictions are well correlated and that none is significantly more accurate than the others at predicting the sense of discomfort.103 All give reasonable predictions for the average discomfort of a group of people but give only poor predictions of an individual's response.104 Given the impediment to trade represented by the different discomfort glare prediction systems in different countries and the apparent lack of difference between them regarding the accuracy of their predictions, the CIE produced a consensus system.105 This system, called the Unified Glare Rating system, uses the formula:

where UGR = Unified Glare Rating, Eb = the illuminance (in lx) on the plane of the eye from the background (excluding the glare source), Li = the luminance (in cd/m2) of the ith part of the glare source in the direction of the eye, ωi = the solid angle (in sr) of the ith part of the glare source, pi = the position index of the ith part of the glare source (consult Chapter 9, Lighting Calculations, for a procedure to calculate position index). This formula results in UGR values ranging from 10 to 30. Different lighting applications can be given a criterion value; for example office lighting is given a limiting value of 20. The relationship (Figure 3-42) between this discomfort glare scale and the more familiar VCP values has been calculated as VCP values of 50%, 60%, 70%, 80%, and 90% correspond to UGR values of 24.0, 21.6, 19.0, 16.0, and 11.6, respectively.106 The accuracy with which the UGR system can predict the level of discomfort produced by a glare source for a group of people has been experimentally tested in the laboratory and in the field and has been found to be high.107 Shadows. Shadows occur when light from a particular direction is intercepted by an opaque object. Large objects reduce the illuminance over a large area. This is typically the problem in industrial lighting where large pieces of machinery cast shadows in adjacent areas. The effect of these shadows can be overcome either by increasing the proportion of interreflected light by using high-reflectance surfaces or by providing local lighting in the shadowed area. If the object is small and close to the area of interest, the shadow can be cast over a meaningful area, which in turn can cause perceptual confusion, particularly if the shadow moves. An example of this is the shadow of a hand cast on a blueprint. This problem can be reduced by increasing the interreflected light in the space or by providing local lighting that can be adjusted in position.

Figure 3-42. The relationship between VCP and the UGR discomfort glare scale. Although shadows can cause visual discomfort, it should be noted that in the form of shading, they are a valuable element in revealing the form of three-dimensional objects. Techniques of display lighting are based around the idea of creating highlights and shadows to change the perceived form of the object being displayed. The number and nature of shadows produced by lighting installations depend on the size and number of light sources and the extent to which light is interreflected. The strongest shadows are produced from a single point source in a black room. Weak shadows are produced when the light sources are large in area and the degree of interreflection is high. Veiling Reflections. Veiling reflections are luminous reflections from specular or semi-matte surfaces that physically change the contrast of the visual task and therefore change the stimulus presented to the visual system. Veiling reflections and disability glare are similar in that both change the luminance contrast of the retinal image but differ in that veiling reflections change the luminance contrast of the task while disability glare changes the luminance contrast of the retinal image. The two factors that determine the nature and magnitude of veiling reflections are the specularity of the material being viewed and the geometry between the observer, the target, and any sources of high luminance. If the object is a perfect diffuse reflector (i.e., a lambertian reflector), no veiling reflections can occur. If it has a specular reflection component, veiling reflections can occur. The positions where veiling reflections occur are those where the incident ray corresponding to the reflected ray that reaches the observer's eye from the target comes from a source of high luminance. This means that the strength and magnitude of such reflections can vary dramatically within a single lighting installation.108 Veiling reflections can cause visual discomfort because they can reduce the luminance contrast and hence the difficulty of the task. In some cases, luminance contrast can increase as veiling reflections increase. One such case is when a combination of specular and diffuse reflecting materials are used for the target and background. Then, a high enough luminance can cause the polarity of luminance contrast to reverse. The effect of veiling reflections on contrast may be quantified by adding the luminance of the veiling reflection to the appropriate components in one of the luminance contrast formulas (Equations 3-6, 3-7, 38). What the appropriate components are depends on the reflection properties of the material being viewed. For glossy ink writing on matte paper, the luminance of the veiling reflections should only be added to the luminance of the ink. For a glossy magazine page or a VDT screen, where there is a specularly reflecting transparent coating over the whole surface, veiling reflections occur over the whole surface. In this case the luminance of the veiling reflections should be added to all terms in the luminance contrast formula. The reflections that occur on materials that have a uniform specularly reflecting surface, such as a VDT

screen, can do more than just modify the luminance contrast. Wherever there is an extended specular reflecting surface, a reflected image of the scene is formed by the surface. This image represents an alternative view of the world to that shown by the printed page or the VDT display. This perceptual conflict can cause discomfort (see "Symptoms and Causes of Visual Discomfort" above). Further, the reflected image is at a different focal distance than the surface, so fluctuations in accommodation and vergence eye movements occur as attention is switched from one image to the other, which may cause fatigue and discomfort to the worker during prolonged viewing. It is therefore quite important to limit the magnitude of veiling reflections in offices (see Chapter 11, Office Lighting). Like shadows, veiling reflections can also be used positively, but in these cases they are called highlights. Physically, veiling reflections and highlights are the same thing. Display lighting of specularly reflecting objects is all about producing highlights to reveal the specular nature of the surface. The extent to which changes in luminance contrast changes visual performance can be estimated using one of the models of visual performance discussed earlier but the extent to which it causes discomfort is different. Whereas veiling reflections affect visual performance when the luminance contrast without veiling reflections is low, it has been shown that about a 20% reduction in luminance contrast is the limit of what is acceptable, regardless of the luminance contrast without veiling reflections (Figure 3-43).109

Figure 3-43. The luminance contrast reduction considered acceptable by 90% of observers plotted against the luminance contrast (see Equation 3-6) of the materials when no veiling reflections occurred.

Comfort, Performance, and Expectations While lighting conditions that compromise visual performance are almost always considered uncomfortable, lighting conditions that allow a high level of visual performance can also be considered uncomfortable. Figure 3-50 below shows the mean detection speed for finding one number from many laid out at random on a table, and the percentage of people considering the lighting "good." As might be expected, increasing the illuminance on the table increases mean detection speed and the percentage considering the lighting "good." However, as the illuminance exceeds 2000 lx (200 fc), the percentage considering the lighting "good" declines even though the mean detection speed continues to increase. Assuming that lower values of "good" means more discomfort, this result indicates both that if you wish to achieve a satisfactory lighting installation it is necessary to provide lighting that allows easy visual performance and avoids discomfort, and that visual discomfort is more sensitive to lighting conditions than visual performance. There is another aspect of visual comfort that distinguishes it from visual performance. Visual performance is determined solely by the capabilities of the visual system. Visual comfort is linked to people's expectations. Any lighting installation that does not meet expectations may be considered uncomfortable even though visual performance is adequate; and expectations can change over time. Figure 3-41 also

suggests another potential impact of visual comfort. Lighting conditions that are considered uncomfortable may influence task performance by changing motivation even when they have no effect on the stimuli presented to the visual system and hence on visual performance.

Approaches to Improving Visual Comfort In order to ensure visual comfort it is necessary to ensure that the lighting allows a good level of visual performance and does not cause distraction. This can be done by the following:

Identify the visual tasks to be performed and then determine the characteristics of the lighting needed to allow a high level of visual performance of the tasks, for example, by using the RVP model. Eliminate flicker from the lighting by using high-frequency control gear for discharge lamps. If this is not possible, reduce the percentage modulation of the perceived flicker by mixing light from sources operating on different phases of the electricity supply. Reduce disability glare by selection, placement, and aiming of luminaires to reduce the luminance of the luminaires close to the common lines of sight. Reduce discomfort glare by selection and layout of luminaires. Use VCP or UGR to estimate the magnitude of discomfort glare. Using high-reflectance surfaces in the space will help reduce discomfort glare by increasing the background luminance against which the luminaires are seen. Consider the density and extent of any shadows that are likely to occur. If shadows are undesirable and large-area shadows are likely to occur, use high-reflectance surfaces in the space to increase the amount of interreflected light and use more lower-wattage lamps to supply the desired illuminance. If shadows cannot be avoided because of the extent of obstruction in the space, provide supplementary task lighting in the shadowed areas. If dense, small-area shadows occur in the immediate work area, use adjustable task lighting to moderate their impact. Reduce veiling reflections by reducing the specular reflectance of the surface being viewed, or by changing the geometry between the viewer, the surface being viewed, and the observer, or by increasing the amount of interreflected light in the space. If the reflections are occurring on a self-luminous surface, such as a VDT screen, use dark letters on a bright background. This will reduce the impact of any veiling reflections seen on the screen.110,111

PERCEPTION OF LIGHTING The perception of the visual world is not solely determined by the physical stimuli presented to the visual system as the retinal image. The existence of a large number of visual illusions is sufficient to demonstrate this. Rather, the stimuli to the visual system provide information which the visual system interprets on the basis of past experience and coincident information. Figure 3-44 shows a surface with dents and dings in it. However, if this page is inverted the dents become dings and vice versa, because it is unconsciously assumed that the light which is casting the shadows always comes from above. When considering how we perceive the world, the overwhelming impression is one of stability in the face of continuous variation. As the eyes move in the head and the head itself moves about, the retinal images of objects move across the retina and change their shape and size according to the laws of physical optics. Further, throughout the day, the spectral emission and distribution of daylight changes as the sun moves across the sky and the meteorological conditions vary. Despite these variations in the retinal image, our perception of reality changes very little. This invariance of perception is called perceptual constancy. The advantage in being able to recognize a tiger as a tiger over a wide range of lighting conditions is obvious.

Figure 3-44. The effect of light on the perception of depth. This surface has both dents (a bulge into the page) and dings (a bulge out of the page). Turning the page upside down makes dents appear as dings and vice-versa because, perceptually, we have learned that the lighting comes from above. Used, by permission, from R. Sekuler and R. Blake, Perception. © 1994. McGrawHill.

The Perceptual Constancies There are four fundamental attributes of an object that are constant over a wide range of lighting conditions. They are: 1. Lightness. Lightness is the perceptual attribute related to the physical quantity of reflectance. In most lighting situations, it is possible to distinguish between the illuminance on a surface and its reflectance, that is, to perceive the difference between a low-reflectance surface receiving a high illuminance and a high-reflectance surface receiving a low illuminance, even when both surfaces have the same luminance. It is this ability to perceptually separate the luminance of the retinal image into its components of illuminance and reflectance that ensures that a piece of coal placed near a window is always seen as black while a piece of paper far from the window is always seen as white, even when the luminance of the coal is higher than the luminance of the paper. 2. Color. Physically, the stimulus a surface presents to the visual system depends on the spectral content of the light illuminating the surface and the spectral reflectance of the surface. However, quite large changes in the spectral content of the illuminant can be made without causing any changes in perceived color. This is evident from the ease with which two similar colors can be discriminated when seen side by side and from the difficulty in discrimination when they are seen successively. See Chapter 4, Color, for additional discussion of color constancies. 3. Size. As an object gets farther away, the size of its retinal image gets smaller but the object itself is not seen as getting smaller. This is because, by using clues such as texture and masking, it is usually possible to estimate the distance and then to compensate unconsciously for the increase in distance. Figure 3-45 shows an illustration of a room, called the Ames room after the inventor, where the cues to distance have been deliberately designed to be misleading when viewed from a specific position. The distortion in perceived size of the people standing in the two corners of the room is startling even after seeing the illusion many times.

4. Shape. As an object changes its orientation in space, its retinal image changes. Nonetheless, in most lighting conditions it is possible to distinguish the orientation in space so a circular plate will always look circular even when its tilted image is elliptical. These constancies represent the application of everyday experience and the integration of all the information about the lighting available in the whole retinal image to the interpretation of a part of the retinal image that bears several alternative interpretations. Given this process it should not be too surprising that the constancies can be broken by restricting the information available coincident with the object being viewed. For example, viewing a surface through an aperture that limits the view to a limited part of the surface will often eliminate lightness constancy. Likewise, eliminating cues to distance will destroy size constancy; changing cues to the plane in which an object is lying will reduce shape constancy and eliminating information on the spectral content of the illuminant will reduce color constancy. In general, constancy is likely to break down whenever there is insufficient or misleading information available from the surrounding parts of the visual field. The conditions recommended for maintaining the constancies are:112

Adequate light No disability glare Good color renderinged High chroma colors, particularly on dimly lighted surfaces A variety of surface colors, including some small white surfaces No large glossy areas Materials with characteristic colors and textures Obvious but not necessarily visible sources of light

Lighting conditions used in display lighting usually set out to break the constancies, particularly lightness constancy.112,113 Even when the lighting conditions are such as to support perceptual constancy, lightness and color constancy will break down if large changes in illuminance or spectral content occur. Figure 3-46 shows the apparent Munsell values of spectrally neutral surfaces, plotted against the illuminance on the surfaces. Figure 3-46 shows that as the illuminance is decreased, the apparent Munsell values (i.e., the lightnesses) are reduced for all the Munsell samples, until at very low illuminance all the Munsell values are in the range of gray to black. It should also be pointed out that this gradual breakdown in lightness constancy requires very large changes in illuminance relative to those that typically occur in interior lighting.

Modes of Appearance While lighting has an important role in preserving or eliminating constancy, it also has a role in determining the perceived visual attributes of objects. It has been argued that objects can have five different attributes-brightness, lightness, hue, saturation, transparency, and glossiness--depending on their nature and the way they are lighted.114 These attributes are defined as follows: 1. Brightness. An attribute based on the extent to which an object is judged to be emitting more or less light. 2. Lightness. An attribute based on the extent to which an object is judged to be reflecting or transmitting a greater or lesser fraction of the incident light. 3. Hue. An attribute based on the classification of a color as reddish, yellowish, greenish, bluish, or their intermediaries, or as having no color. 4. Saturation. An attribute based on the extent to which a color is different from a color of the same brightness or lightness. 5. Transparency. An attribute based on the extent to which colors are seen behind or within an

object. 6. Glossiness. An attribute based on the extent to which a surface is different from a matte surface with the same lightness, hue, saturation, and transparency.

Figure 3-45. The Ames room: a demonstration that providing false cues to distance will break size constancy.

Figure 3-46. Apparent Munsell values at different illuminances for surfaces seen against a background of reflectance = 0.2. The vertical line at an illuminance of 786 lx indicates the reference condition. At this illuminance the apparent Munsell values of the surfaces have been normalized to their actual Munsell values. Not all these attributes occur in every situation. Rather, different combinations of attributes occur in different modes of appearance. The four modes of appearance are: 1. Aperture mode. This occurs when an object or surface has no definite location in space, as occurs when a surface is viewed through an opening. 2. Illuminant mode. This occurs when an object or surface is seen to be emitting light. 3. Object mode, volume. This occurs when a three-dimensional object has a definite location in space with defined boundaries. 4. Object mode, surface. This occurs when a two-dimensional surface has a definite location in space with defined boundaries.

The following table shows which of the attributes are associated with each mode of appearance.

Of particular interest to the perception of lighting is the shift between the attributes of brightness and lightness in different modes of appearance. An object that appears in the self-luminous mode, such as a VDT screen or a lamp, is perceived to have a brightness but not a lightness. In this mode of appearance, the concept of reflectance is meaningless. However, an object that appears in the volume mode, such as a VDT screen or a lamp that is turned off, does not have an attribute of brightness but does have a lightness in that its reflectance can be estimated. A similar transformation occurs between the volume or surface modes of appearance and the aperture mode. Even reflective objects, when seen in the aperture mode, are perceived as having a brightness but not a lightness. When seen in the object mode they have a lightness and not a brightness. This is important because lighting can be used to change the mode of appearance. For example, a hung painting has a lightness attribute when lighted so that both it and the wall appear in the surface mode. However, if the painting is illuminated solely with a carefully aimed framing spot so that the edge of the beam coincides with the edges of the painting, the painting is seen in the aperture mode and takes on a self-luminous quality with a brightness attribute. Adjusting the modes of appearance is an important technique in display lighting, both indoors and outdoors.

Brightness Perception For objects in the illuminant or aperture modes, the perception of brightness is a function of luminance. Specifically, brightness is related to luminance by a power law115 of the form:

where B = brightness, a = constant, L = luminance in cd/m2. Despite the logical inconsistency of describing a perception of brightness to an object in the surface mode, studies of the perception of brightness of surfaces in an interior have been made.116 The room contained both self-luminous objects, which appeared in the illuminant mode, and reflecting surfaces, which appeared in the surface mode. The luminance range in the interior was two log units. These studies have shown that the perceived brightness of any single surface increases with luminance according to a power law with an exponent of 0.35, but that the brightness of a number of surfaces seen simultaneously follows a power law with an exponent of approximately 0.6. These relationships can be used to estimate the relative brightness of surfaces in an interior by assuming that the brightest surface in the room has a brightness given by:

Then, other surfaces will have a brightness given by:

This simple system underestimates the brightness of highly saturated colored surfaces and overestimates the brightness of translucent surfaces. These relationships are given for guidance only. The data from which these relationships were derived represented just over half the data collected. The data from other subjects were eliminated to reduce the "noise" in the data, "noise" that may reflect the lack of meaning some people find in attempting to describe the brightness of an object in the surface mode.

Holistic Perceptions The above discussion of perceptual constancy and modes of appearance is concerned mainly with the perception of individual objects in an interior. This section is concerned with the factors that determine the perception of the whole interior. Lighting Cues. An experiment designed to address the holistic perception of a space was carried out in a small office lit in eighteen different ways.117 The illuminance on the desk was always 500 lx, but the distribution of light in the rest of the room varied widely. A panel of people was asked to evaluate the room lighted by each lighting installation using an extensive questionnaire. Using factor analysis, three independent factors were identified in the responses: the first was simply whether people liked the installation; the second was the brightness, in the sense of the amount of light in the office; the third was interest. Figure 3-47 shows the positions of the eighteen installations (a) on a map formed by the brightness and interest dimensions. The contours on the map show the relative preference for the different parts of the map. Clearly, for this work space people preferred lighting that was both bright and interesting. Regular arrays of luminaires can produce brightness but are rarely interesting. Irregular arrangements of lighting equipment can be interesting but may not produce enough brightness. Designing lighting to be bright and interesting would seem to be a good approach to designing lighting for work spaces.

Figure 3-47a. A map showing the location of the listed office lighting installations on the two

dimensions, interest and brightness, identified by factor analysis. Superimposed on this map are isopreference contours based on preference ratings of the same installations. These contours define areas of equal preference from area A (most preferred) to area E (least preferred).

Figure 3-47b. Installations Used by Hawkes et al.

Figure 3-48. A Summary of Lighting Cues to Produce Specific Impressions Behind this approach lies the belief that the experience of room lighting is, in part, an experience of interpreting complex light patterns. This implies that different patterns of light can be treated as cues to the "meaning" of the space, which in turn carries information about its likely suitability for its function. The most comprehensive data on how perceptions of space can be altered by changing the amount and distribution of light come from the work of Flynn and his colleagues.118,119 Following an extensive series of experiments, Flynn and his colleagues identified a series of lighting cues that could be used to reinforce specific perceptions. Figure 3-48 sets out these cues. Some designers have found this listing to be useful in their work. Various aspects of the methodology used in some of these studies have been criticized,120,121 so the guidelines given in Figure 3-48 should be treated as indicative rather than definitive. Nonetheless, there is sufficient evidence to suggest that the underlying concept--that different amounts and distributions of light can change the perception of a space--is correct. Gloom. An alternative methodology has been used to explore the perception of gloom in obviously functional spaces.122,123 Gloom is likely to be perceived when any of the following conditions are provided:

Low surround luminances, irrespective of task illuminance Conditions in which fine detail in the periphery are obscured High task illuminances with low luminances on the peripheral surfaces Adaptation luminances in the mesopic region.

Room Brightness. Given that the surfaces in a conventionally lighted room appear in the object mode, it seems likely that when one talks about the brightness of a room the amount of light in the room is being evaluated. The most obvious lighting variable determining the perception of room brightness is the illuminance on the working plane.118,119 However, there is evidence that maximum luminance, light distribution, and light spectrum also influence the perception of room brightness to a significant extent. Studies have shown that the presence of a sparkling luminous element increases the perceived brightness of

the room by approximately 20%.124,125 As for light distribution, a study of room brightness for a uniformly and a nonuniformly illuminated room, using a psychophysical technique, showed that the latter required about 5 to 10% less illuminance on the working plane to match the uniformly lit room for equal room brightness.126 Finally, the effect of light spectrum on perceived room brightness has been studied for many years under the title of visual clarity. Several studies have shown that light spectra that give surface colors greater saturation require about 10 to 30% lower working plane illuminances for the rooms to be seen as equally bright.127-129 Why this should occur is unknown, but it has been suggested that it has to do with the scotopic content of the spectrum.130 The brightness perception appears to be related to the function of rods, even at photopic light levels. These results are consistent with the other holistic studies in that there are several different ways to achieve the same perception. Further, the perception is influenced by the lighting of the whole space, not just the task area. It is important to note that these holistic studies have all taken place in functional interiors. It is at least possible that the holistic perception generated by the same lighting conditions varies with context. For example, lighting conditions that generate a perception of gloom in an office might generate a perception of privacy in a restaurant.

Guidelines for Acceptable Lighting There have been a number of studies that have examined the acceptability of specific aspects of lighting, usually in a specific context. This section considers what is known about the preference for three lighting variables, all of which are under the control of the lighting designer. Illuminance. The acceptability of different illuminances in offices and other working areas has been examined to try to determine the effect of illuminance on observer preference.131 Figure 3-49 shows the percentage of observers considering the lighting of a room with systematically varied room surface luminances up to 150 cd/m2 to be too dark, good, or too bright. Independent of the wall and ceiling luminances, the maximum proportion of "good" appraisals occurred when the luminance of the working plane was 130 cd/m2.

Figure 3-49. The percentages of observers rating the luminance of their desks as too dark, good, or too bright. These findings were verified and extended over a wider range of illuminance.132-134 In one experiment,132 the task of searching a random array of numbers for a particular number was carried out under illuminances ranging from 50 to 10,000 lx, and with varying contrast (black numbers on white and gray paper). Each subject gave an opinion of the lighting, indicating whether it was "too dark," "good," or "too bright." Judgments of optimum illuminance increased with age and with decreasing task contrast. Most subjects preferred 1000 lx when searching the higher-contrast number lists, but preferred 1800 lx when searching the lower-contrast materials. On average, for both contrast levels, the younger subjects (less than 50 years of

age) indicated that 2000 lx was preferable, while the older subjects required 5000 lx to achieve comparable satisfaction. At the highest illuminances (5000 to 10,000 lx), rated acceptability decreased even though performance of the task continued to increase (Figure 3-50).

Figure 3-50. (a) Mean detection speeds for locating a specified number from among others at different illuminances; (b) the percentage who considered the lighting "good," at each illuminance

Figure 3-51. Preferred Illuminances in Offices

Figure 3-52. Preferred Luminance Ratios of Room Surfaces This general trend of increased preference for higher illuminances for more visually difficult tasks, followed by a decrease in satisfaction at even higher illuminances, has been replicated by subsequent investigators using a variety of tasks and subjective scaling techniques in offices.71,135-138 Ranges of preferred illuminances identified in these studies are presented in Figure 3-51.131-133,137-139 An inspection of Figure 3-51 suggests that the illuminance ranges are wide and that there is no sharp preference for a specific illuminance. This is to be expected from the plateau of visual performance discussed in "Models of On-Axis Visual Performance." Of course, the illuminance ranges do reflect the effect of different tasks and the wide range of individual differences present in any subjective judgment of lighting. These individual differences are the most likely reason why the percentage of people considering the lighting good in Figure 3-49 does not exceed approximately 70%. Spatial Distribution. Reports that changing the spatial distribution of light affects vision have been known since the mid-nineteenth century.41 The effects of different spatial distributions of light on visual performance and preference have been studied extensively. Many studies have focused on performance.41,140-149 In general, the more uniform the light distribution in the visual field and the larger the area of the visual field it covers, the better one sees the visual task. The spatial distribution of illuminance across the working area is also important. Studies of acceptable illuminance uniformity across an office desk have shown that acceptability starts to decline as the minimum-to-maximum illuminance ratio over the working area falls below approximately 0.7.138,149 It should be noted that this nonuniformity consisted of a steady change in illuminance across the desk such as would be produced by large spacings in a regular array of luminaires. If the variation in illuminance had involved a sharp edge, such as would be produced by the shadow of a shelf, acceptability would start to decline at a higher minimum-to-maximum illuminance ratio. Another aspect of localized illuminance distributions is the preferred ratio of task to desktop luminances under different levels of ambient illuminance. Subjects were asked to sit at each of six desks (different unspecified reflectance of each desk top) under four illuminances (50, 100, 500, 1000 lx) and copy figures from one white sheet of paper onto another.150 They were then asked to indicate at which desk they preferred to perform this task under the different illuminances. As the illuminance increased, the subjects preferred lower-reflectance desk tops. For higher illuminances (500 lx) the preferred ratio between the paper and the desk was 3:1, whereas for lower illuminances 2:1 was preferred. Subsequent investigators broadened the scope of this work, examining the effects on preferences of varying the luminances of surfaces other than the immediate surround, such as walls and ceilings. Figure 3-52 summarizes the results of these studies.139,150-153 Inspection of Figure 3-52 suggests that although some consensus on preferred ratios of task to immediate surround luminances can be identified, less agreement exists about preferences for luminances of more remote surfaces. Obviously, more systematic research is required before a complete specification of preferred luminance ratios throughout the visual environment can be identified.

As a guide for design purposes, luminance ratio limits have been recommended for various applications, such as offices, educational facilities, institutions, industrial areas, and residences (see the application chapters). For additional guidance, recommended limits on reflectances (both upper and lower) of large surfaces are given for the same applications. The use of these reflectance limits, along with a selection of appropriate colors, should help to control luminances and keep within the ratio limits without creating a bland and uninteresting environment. Color of Illumination. The color of illumination can be described by two independent properties: chromaticity, or correlated color temperature (CCT), and color rendering. There is often confusion between chromaticity and color rendering. In simple terms, chromaticity refers to the color appearance of a light source, "warm" for low CCT values and "cool" for high CCT values. Color rendering refers to the ability of a light source, with its particular CCT, to render the colors of objects the same as a reference light source of the same CCT. This aspect is typically measured in terms of the CIE General Color Rendering Index. For more information see Chapter 4, Color.

Figure 3-53. The Kruithof effect: The white area defines the preferred combinations of the color temperature of a light source and the illuminance. Color temperature/illuminance combinations in the lower shaded area are claimed to produce cold, drab environments, while those in the upper shaded area are believed to produce overly colorful and unnatural environments. Experiments examining the psychological effects of varying CCT and illuminance have suggested that using lamps with high CCT values at low illuminances will make a space appear cold and dim. Conversely, using lamps with low CCT values at high illuminances will make a space appear artificial and overly colorful. Figure 3-53 illustrates this so-called Kruithof effect.154 Although these findings have been broadly replicated,155 other investigators have failed to find a similar tradeoff of CCT and illuminance.156,157 Rather they found that when people spent sufficient time in the room for color adaptation to occur, the perceptions of rooms lighted with lamps of different color temperature was dominated by illuminance. This implies that where color adaptation occurs with no opportunity to compare lamps with different CCTs the CCT of the light source is relatively unimportant to perception. Where comparisons can be made or color adaptation does not occur, CCT is more likely to be important. At the very least this confusion means that the widespread belief about the tradeoff of CCT and illuminance should be treated with some skepticism.

LIGHTING AND BEHAVIOR Lighting has been shown to have an effect on overt behavior. Specifically, lighting can be used to cue orientation and circulation in humans and increases alertness and activation. However, these findings should

be interpreted with caution for at least two reasons. First, the range of illuminances studied has been small. In many experiments the effects of only two or three levels have been studied. More important, the effects of illuminance have been shown to interact with other independent variables.158,159 Simple models are often insufficient for describing relationships among different environmental variables; effects may be facilitated or inhibited depending on the presence of other moderating factors.160 Hence, further studies using a wider range of lighting conditions and additional independent variables are required before firm conclusions about the effects of illuminance on orientation, wayfinding, activation, and attention can be drawn. Light clearly affects spatial orientation and wayfinding. For example, when navigating around a barrier, people tend to follow the direction of higher illuminance.161 These results support the notion that the distribution of light might be used to direct circulation, and as an aid to wayfinding. Similar findings have been reported in another context. An investigation of the effects of spatial distribution of light on seat choice and orientation in a cafeteria showed that people selected seats facing bright areas. When the lighting was changed to highlight a different surface, patterns of seat selection and orientation changed to face the new bright area.118 The effects of wall lighting on desk selection have also been observed. Subjects entered a room and sat at one of three desks to complete a series of questionnaires. Desks were located next to the door, in the middle of the room, and at the far side of the room opposite the door. When the wall opposite the door was illuminated, most subjects crossed the room and sat at the desk located next to that wall. When that wall was not illuminated, most subjects sat at the desk located next to the door.162 These studies can be interpreted as examining the use of light as a means of directing attention. A more direct study of this possibility involved supplementary classroom lighting in a primary school.163 Lists of words used in spelling tests were displayed at the front of classrooms. Supplementary lighting was used to highlight the word lists in one condition of the experiment but not in the control condition. Significantly more inattentive behaviors were coded in the control condition than when the word lists were highlighted. Lighting can also affect activities not directly related to vision. A significant reduction in sound level in a school hallway was found when the illuminance was low.164 Other researchers extended these findings by examining possible interactions between illuminances and other variables in their effects on human performance. An interaction was found between light and sound levels in their effects on the performance of a reaction time task. The presence of a white-noise sound increased reaction times under higher illuminances and had no effect in the dark.165 More generally, when working at night, exposure to bright light has been shown to increase core body temperature and brain activity in ways usually associated with increased alertness166 and to change performance on complex cognitive tasks.167 Whether similar alerting effects of light exposure occur during daytime remains to be determined.

REFERENCES 1. Sekuler, R., and R. Blake. 1994. Perception. New York, NY: McGraw-Hill. 2. Boettner, E. A., and J. R. Wolter. 1962. Transmission of the ocular media. Invest. Opthalmol. 1(6):776783. 3. Said, F. S., and R. A. Weale. 1959. The variation with age of the spectral transmissivity of the living human crystalline lens. Gerontologia 3(4):213-231. 4. Brainard, G. C., M. D. Rollag, and J. P. Hanifin. 1997. Photic regulation of melatonin in humans: Ocular and neural signal transduction. J. Biolog. Rhythms 12(6):537-546. 5. Coren, S., and J. S. Girgus. 1972. Density of human lens pigmentation: In vivo measures over an extended age range [Letter]. Vision Res. 12(2):343-346.

6. Vos, J. J., and J. Boogaard. 1963. Contribution of the cornea to entoptic scatter. J. Opt. Soc. Am. 53 (7):869-873. 7. Boynton, R. M., and F. J. J. Clarke. 1964. Sources of entoptic scatter in the human eye. J. Opt. Soc. Am. 54(1):110-119. 8. Wyszecki, G., and W. S. Stiles. 1982. Color science: Concepts and methods, quantitative data and formulae. 2nd ed. New York: John Wiley & Sons. 9. Vos, J. J. 1963. Contribution of the fundus oculi to entoptic scatter. J. Opt. Soc. Am. 53(12):1449-1451. 10. Wolf, E., and J. S. Gardiner. 1965. Studies on the scatter of light in the dioptric media of the eye as a basis of visual glare. Arch. Ophthalmol. 74(3):338-345. 11. Weale, R. A. 1985. Human lenticular fluorescence and transmissivity, and their effects on vision. Exp. Eye Res. 41(4): 457-473. 12. Krueger, H. 1991. Visual function and monitor use. In The man-machine interface, edited by J. A. Roufs. Vision and Visual Dysfunction, volume 15. London: MacMillan Press. 13. Leibowitz, H. W., and D. A. Owen. 1975. Anomalous myopias and the intermediate dark focus of accommodation. Science 189(4203):646-648. 14. Kaiser, P. K., and R. M. Boynton. 1996. Human color vision. Washington: Optical Society of America. 15. Alman, D. H. 1977. Errors of the standard photometric system when measuring the brightness of general illumination light sources. J. Illum. Eng. Soc. 7(1):55-62. 16. Ingling, C. R. Jr., and H. B. Tsou. 1997. Orthogonal combinations of three visual channels. Vision Res. 17(9):10751082. 17. Bouma, H. 1965. Receptive systems mediating certain light reactions of the pupil of the human eye. Philips Research Report Supplements, no. 5. Eindhoven, Netherlands: Philips Research Laboratories. 18. Weale, R. A. 1992. The senescence of human vision. New York: Oxford University Press. 19. Dowling, J. A. 1967. The site of visual adaptation. Science 155(3760):273-279. 20. Hecht, S., and J. Mandelbaum. 1939. The relation between vitamin A and dark adaptation. JAMA 112 (19):1910-1916. 21. Boynton, R. M., and N. D. Miller. 1963. Visual performance under conditions of transient adaptation. Illum. Eng. 58(8): 541-550. 22. He, Y., M. Rea, A. Bierman, and J. Bullough. 1997. Evaluating light source efficacy under mesopic conditions using reaction times. J. Illum. Eng. Soc. 26(1):125-138. 23. Commission Internationale de l'Éclairage. 1989. Mesopic Photometry: History, special problems and practical solutions. CIE no. 81. Vienna: Bureau Central de la CIE. 24. Kaiser, P. K., and G. Wyszecki. 1978. Additivity failures in heterochromatic brightness matching. Color Res. Appl. 3(4): 177-182. 25. Wagner, G., and R. M. Boynton. 1972. Comparison of four methods of heterochromatic photometry. J. Opt. Soc. Am. 62(12):1508-1515.

26. Guth, S. L., and H. R. Lodge. 1973. Heterochromatic additivity, foveal spectral sensitivity, and a new color model. J. Opt. Soc. Am. 63(4):450-462. 27. He, Y., A. Bierman, and M. S. Rea. 1998. A system of mesopic photometry. Light. Res. Tech. 30 (4):175-181. 28. Judd, D. B. 1951. Report of the U.S. Secretariat Committee on colorimetry and artificial daylight. Proceedings of the Commission Internationale de l'Éclairage, 12th Session, Paris: Bureau Central de la CIE. 29. Commission Internationale de l'Éclairage. 1978. Light as a true visual quantity: Principles of measurement. CIE Publication no. 41. Vienna: Bureau Central de la CIE. 30. Gibson, K. S., and E. P. T. Tyndall. 1923. Visibility of radiant energy. Bulletin Bureau of Standards 19:131. 31. Commission Internationale de l'Éclairage. 1994. CIE Technical Report No. 107: Review of the official recommendations of the CIE for the colors of signal lights. Vienna: Bureau Central de la CIE. 32. Sekuler, R., D. Kline, and K. Dismukes, Eds. 1982. Aging and human visual function. Modern Aging Research, 2. New York: Alan R. Liss, Inc. 33. Blackwell, O. M., and H. R. Blackwell. 1971. Visual performance data for 156 normal observers of various ages. J. Illum. Eng. Soc. 1(1):3-13. 34. Sorensen, S., and G. Brunnstrom. 1995. Quality of light and quality of life: An intervention study among older people. Light. Res. Tech. 27(2):113-118. 35. Kahn, H. A. 1973. Statistics on blindness in the model reporting area 1969-1970. Department of Health, Education and Welfare no. 73-427. Washington: National Institutes of Health. 36. Cullinan, T. R. 1977. The epidemiology of visual disabilities studies of visually disabled people in the community. Canterbury: University of Kent. 37. Commission Internationale de l'Éclairage. 1997. Low vision: Lighting needs for the partially sighted. CIE Publication no. 123. Vienna: Bureau Central de la CIE. 38. Sicurella, V. J. 1977. Color contrast as an aid for visually impaired persons. JVIB 71(6):252-257. 39. Blackwell, H. R. 1946. Contrast thresholds of the human eye. J. Opt. Soc. Am. 36(11):624-643. 40. Boff, K. R., and J. E. Lincoln. 1988. Engineering data compendium: Human perception and performance. Wright-Patterson Air Force Base, Ohio: Harry G. Armstrong Aerospace Medical Research Laboratory. 41. Lythgoe, R. J. 1932. The measurement of visual acuity. Medical Research Council Special Report, No. 173. London: H.M. Stationary Office. 42. Blackwell, H. R., and O. M. Blackwell. 1980. Population data for 140 normal 20-30 year olds for use in assessing some effects of lighting upon visual performance. J. Illum. Eng. Soc. 9(3):158-174. 43. Nadler, M. P., D. Miller, and D. J. Nadler. 1990. Glare and contrast sensitivity for clinicians. New York: Springer-Verlag. 44. Berman, S. M., D. S. Greenhouse, I. L. Bailey, R. D. Clear, and T. W. Raasch. Human electroretinogram responses to video displays, fluorescent lighting, and other high frequency sources. Opt.

Vis. Sci. 68(8):645-662. 45. Bedford, R. E., and G. W. Wyszecki. 1958. Wavelength discrimination for point sources. J. Opt. Soc. Am. 48(2): 129-135. 46. Wright, W. D. 1946. Researches on normal and defective color vision. London: Henry Kimpton. 47. Robertson, A. R. 1981. Color differences. Die Farbe 29:273. 48. MacAdam, D. L. 1942. Visual sensitivities to color differences in daylight. J. Opt. Soc. Am. 32(5):247274. 49. Roethlisberger, F. J., and W. J. Dickson. 1934. Management and the worker: Technical vs. social organization in an industrial plant. Boston: Harvard University Press. 50. Smith, S. W., and M. S. Rea. 1978. Proofreading under different levels of Illumination. J. Illum. Eng. Soc. 8(1):47-52. 51. Smith, S. W., and M. S. Rea. 1980. Relationships between office task performance and ratings of feelings and task evaluations under different light sources and levels. Proceedings: 19th Session, Commission Internationale de l'Éclairage. Paris: Bureau Central de la CIE. 52. Smith, S. W., and M. S. Rea. 1982. Performance of a reading test under different levels of illumination. J. Illum. Eng. Soc. 12(1):29-33. 53. Smith, S. W., and M. S. Rea. 1987. Check value verification under different levels of illumination. J. Illum. Eng. Soc. 16(1):143-149. 54. Blackwell, H. R. 1959. Development and use of a quantitative method for specification of interior illumination levels on the basis of performance data. Illum. Eng. 54(6):317-353. 55. Commission Internationale de l'Éclairage Technical Committee TC-3.1. 1972. A Unified Framework of Methods for Evaluating Visual Performance Aspects of Lighting. Publication CIE no. 19. Paris: Bureau Central de la CIE. 56. Commission Internationale de l'Éclairage. 1981. An analytic model for describing the influence of lighting parameters upon visual performance, Volume 1: Technical foundations. CIE Publication no. 19/2.1. Paris: Bureau Central de la CIE. 57. Rea, M. S. 1983. The validity of the relative contrast sensitivity function for modelling threshold and suprathreshold responses. The Integration of Visual Performance Criteria into the Illumination Design Process, Ottawa: Public Works Canada. 58. Weston, H. C. 1935. The relation between illumination and visual efficiency: The effect of size of work. Prepared for Industrial Health Research Board (Great Britain), and Medical Research Council (London). London: H. M. Stationery Office. 59. Weston, H. C. 1945. The relation between illumination and visual efficiency: The effect of brightness contrast. (Great Britain) and Medical Research Council (London). Industrial Health Research Board Report no. 87. London: H. M. Stationery Office. 60. Rea, M. S. 1987. Toward a model of visual performance: A review of methodologies. J. Illum. Eng. Soc. 16(1):128-142. 61. Rea, M. S. 1981. Visual performance with realistic methods of changing contrast. J. Illum. Eng. Soc. 10 (3):164-177.

62. Rea, M. S. 1986. Toward a model of visual performance: Foundations and data. J. Illum. Eng. Soc. 15 (2):41-57. 63. Boyce, P. R., and M. S. Rea. 1987. Plateau and escarpment: The shape of visual performance. Proceedings: 21st session, Commission Internationale de l'Éclairage. Paris: Bureau Central de la CIE. 64. Rea, M. S., and M. J. Ouellette. 1988. Visual performance using reaction times. Light. Res. Tech. 20 (4):139-153. 65. Rea, M. S., and M. J. Ouellette. 1991. Relative visual performance: A basis for application. Light. Res. Tech. 23(3):135144. 66. Bailey, I., R. Clear, and S. Berman. 1993. Size as a determinant of reading speed. J. Illum. Eng. Soc. 22 (2):102-117. 67. McNelis, J. F. 1973. Human performance: A pilot study. J. Illum. Eng. Soc. 2(3):190-196. 68. Clear, R. 1996. Relationships between the VL and reaction time models. J. Illum. Eng. Soc. 25(2):1424. 69. Clear, R., and R. G. Mistrick. 1996. Multilayer polarizers: A review of the claims. J. Illum. Eng. Soc. 25 (2):70-88. 70. Inditsky, B., H. W. Bodmann, and H. J. Fleck. 1982. Elements of visual performance: Contrast metric-visibility lobes--eye movements. Light. Res. Tech. 14(4):218-231. 71. Smith, S. W., and M. S. Rea. 1980. Relationships between office task performance and ratings of feelings and task evaluations under different light sources and levels. Proceedings: 19th session, Commission Internationale de l'Éclairage. Paris: Bureau Central de la CIE. 72. Williams, L. G. 1966. The effect of target specification on objects fixated during visual search. Perc. Psyc. 1(9):315-318. 73. National Research Council Canada. 1994. Full-spectrum lighting effects on performance, mood, and health, edited by Jennifer A. Veitch, Institute for Research in Construction Report no. 659. Ottawa: National Research Council Canada. 74. Berman, S. M., G. Fein, D. L. Jewett, and F. Ashford. 1993. Luminance-controlled pupil size affects Landolt C task performance. J. Illum. Eng. Soc. 22(2):150-165. 75. Berman, S. M., G. Fein, D. L. Jewett, and F. Ashford. 1994. Landolt-C recognition in elderly subjects is affected by scotopic intensity of surround illuminants. J. Illum. Eng. Soc. 23(2):123-130. 76. Berman, S., G. Fein, D. Jewett, B. Benson, T. Law, and A. Myers. 1996. Luminance-controlled pupil size affects word-reading accuracy. J. Illum. Eng. Soc. 25(1):51-59. 77. Berman, S. M., G. Fein, D. L. Jewett, G. Saika, and F. Ashford. 1992. Spectral determinants of steadystate pupil size with full field of view. J. Illum. Eng. Soc. 21(2):3-13. 78. Wilkins, A. J. 1995. Visual Stress. Oxford: Oxford University Press. 79. Kaplan, S., and R. Kaplan. 1982. Environment and cognition: Functioning in an uncertain world. Ann Arbor: Ulrich's. 80. Rea, M. S., and M. J. Ouellette. 1988. Table-tennis under high intensity discharge (HID) lighting. J.

Illum. Eng. Soc. 17(1):29-35. 81. Wilkins, A. J., I. Nimmo-Smith, A. I. Slater, and L. Bedocs. 1989. Fluorescent lighting, headaches and eyestrain. Light. Res. Tech. 21(1):11-18. 82. Fry, G. A. 1954. A re-evaluation of the scattering theory of glare. Illum. Eng. 49(2):98-102. 83. Holladay, L. L. 1926. The fundamentals of glare and visibility. J. Opt. Soc. Am. 12(4):271-319. 84. Holladay, L. L. 1927. Action of a light source in the field of view on lowering visibility. J. Opt. Soc. Am. 14(1):1-15. 85. Stiles, W. S. 1929. The effect of glare on the brightness difference threshold. Proc. R. Soc. Lond. Ser. B 104(731): 322-351. 86. Fugate, J. M., and G. A. Fry. 1956. Relation of changes in pupil size to visual discomfort. Illum. Eng. 51 (7):537-549. 87. Fry, G. A., and V. M. King. 1975. The pupillary response and discomfort glare. J. Illum. Eng. Soc. 4 (4):307-324. 88. Berman, S. M., R. J. Jacobs, M. A. Bullimore, L. L. Bailey, N. Ghandi, and D. S. Greenhouse. 1991. An objective measure of discomfort glare. First International Symposium on Glare. New York: Lighting Research Institute. 89. Fischer, D. 1991. Discomfort glare in interiors. First International Symposium on Glare, New York: Lighting Research Institute. 90. Luckiesh, M., and S. K. Guth. 1949. Brightness in visual field at borderline between comfort and discomfort (BCD). Illum. Eng. 44(11):650-670. 91. Illuminating Engineering Society. Committee on Recommendations for Quality and Quantity of Illumination. Subcommittee on Direct Glare. 1966. Outline of a standard procedure for computing visual comfort ratings for interior lighting: Report No. 2. Illum. Eng. 61(10):643-666. 92. Hopkinson, R. G. 1957. Evaluation of glare. Illum. Eng. 52(6):305-316. 93. Guth, S. K., and J. F. McNelis. 1959. A discomfort glare evaluator. Illum. Eng. 54(6):398-406. 94. Guth, S. K., and J. F. McNelis. 1961. Further data on discomfort glare from multiple sources. Illum. Eng. 56(1):46-57. 95. Bradley, R. D., and H. L. Logan. 1964. A uniform method for computing the probability of comfort response in a visual field. Illum. Eng. 59(3):189-206. 96. Guth, S. K. 1963. A method for the evaluation of discomfort glare. Illum. Eng. 57(5):351-364. 97. Allphin, W. 1966. Influence of sight line on BCD judgments of direct discomfort glare. Illum. Eng. 61 (10):629-633. 98. Allphin, W. 1968. Further studies of sight line and direct discomfort glare. Illum. Eng. 63(1):26-31. 99. Levin, R. E. 1973. An evaluation of VCP calculations. J. Illum. Eng. Soc. 2(4):355-361. 100. Illuminating Engineering Society. Committee on Recommendations for Quality and Quantity of Illumination. 1972. An alternate simplified method for determining the acceptability of a luminaire from the

VCP standpoint for use in large rooms: RQQ Report no. 3. J. Illum. Eng. Soc. 1(3): 256-260. 101. Fry, G. A. 1976. A simplified formula for discomfort glare. J. Illum. Eng. Soc. 8(1):10-20. 102. Goodbar, I. 1976. A simplified method for determining the acceptability of a luminaire from the VCP standpoint. J. Illum. Eng. Soc. 8(1):21-28. 103. H. Manabe. 1976. The assessment of discomfort glare in practical lighting situations. Oteman Economic Studies no.9. Osaka: Oteman Gakuin University. 104. Boyce, P. R., V. H. C. Crisp, R. H. Simons, and E. Rowlands. 1980. Discomfort glare sensation and prediction. Proceedings: 19th Session. Commission Internationale de l'Éclairage. Paris: Bureau Central la CIE. 105. Commission Internationale de l'Éclairage. 1995. Discomfort glare in interior lighting. CIE Publication no. 117. Vienna: Bureau Central de la CIE. 106. Sorensen, K. 1991. Practical aspects of discomfort glare evaluation: Interior lighting. First International Symposium on Glare. New York: Lighting Research Institute. 107. Akashi, Y., R. Muramatsu, and S. Kanaya. 1996. Unified Glare Rating (UGR) and subjective appraisal of discomfort glare. Light. Res. Tech. 28(4):199-206. 108. Boyce, P. R. 1978. Variability of contrast rendering factor in lighting installations. Light. Res. Tech. 10 (2):94-105. 109. Bjorset, H. H. 1979. A proposal for recommendations for the limitation of the contrast reduction in office lighting. Proceedings of the Commission Internationale de l'Éclairage, 19th Session. Vienna: Bureau Central de la CIE. 110. Boyce, P. R. 1991. Lighting and lighting conditions. In The man-machine interface, Edited by A. J. Roufs. Vision and Visual Dysfunction vol 15. London: Macmillan Press. 111. Lloyd, C. J., M. Mizukami, and P. R. Boyce. 1996. A preliminary model of lighting-display interaction. J. Illum. Eng. Soc. 25(2):59-69. 112. Lynes, J. A. 1994. Daylight and the appearance of indoor surfaces. Proceedings of the CIBSE National Lighting Conference. London: Chartered Institution of Building Services Engineers. 113. Lynes, J. A. 1971. Lightness, colour and constancy in lighting design. Light. Res. Tech. 3(1):98-110. 114. Judd, D. B. 1961. A five-attribute system of describing visual appearance. ASTM Special Technical Publication 297. Philadelphia: American Society for Testing Materials. 115. Stevens, S. S. 1960. Psychophysics of sensory function. American Scientist. 48(2):226-252. 116. Marsden, A. M. 1970. Brightness-luminance relationships in an interior. Light. Res. Tech. 2(1):10-16. 117. Hawkes, R. J., D. L. Loe, and E. Rowlands. 1979. A note towards the understanding of lighting quality. J. Illum. Eng. Soc. 8(2):111-120. 118. Flynn, J. E. 1977. A study of subjective responses to low energy and nonuniform lighting systems. Light. Des. Appl. 7(2): 6-15. 119. Flynn, J. E., and G. J. Subisak. 1978. A procedure for qualitative study of light level variations and

system performance. J. Illum. Eng. Soc. 8(1):28-35. 120. Tiller, D. K. 1990. Toward a deeper understanding of psychological aspects of lighting. J. Illum. Eng. Soc. 19(2):59-65. 121. Veitch, J. A., and G. R. Newsham. 1998. Determinants of lighting quality I: State of the science. J. Illum. Eng. Soc. 27(1):92-106. 122. Shepherd, A. J., W. G. Julian, and A. T. Purcell. 1989. Gloom as a psychophysical phenomenon. Light. Res. Tech. 21(3): 89-97. 123. Shepherd, A. J., W. G. Julian, and A. T. Purcell. 1992. Measuring appearance: Parameters indicated from gloom studies. Light. Res. Tech. 24(4):203-214. 124. Bernecker, C. A., and J. M. Mier. 1985. The effect of source luminance on the perception of environment brightness. J. Illum. Eng. Soc. 15(1):253-271. 125. Akashi, Y., I. Akashi, Y. Tanabe, and S. Kanaya. 1995. The sparkle effect of luminaires on the sensation of brightness. Proceedings of the Commission Internationale de l'Éclairage, 23rd session. Vienna: Bureau Central de la CIE. 126. Tiller, D. K., and J. A. Veitch. 1995. Perceived room brightness: Pilot study on the effect of luminance distribution. Light. Res. Tech. 27(2):93-101. 127. Aston, S. M., and H. E. Bellchambers. 1969. Illumination, colour rendering and visual clarity. Light. Res. Tech. 1(4):259-261. 128. Bellchambers, H. E., and A. C. Godby. 1972. Illumination, color rendering and visual clarity. Light. Res. Tech. 4(2): 104-116. 129. Boyce, P. R. 1977. Investigations of the subjective balance between illuminance and lamp colour properties. Light. Res. Tech. 9(1):11-24. 130. Berman, S. M., D. L. Jewett, G. Fein, G. Saika, and F. Ashford. 1990. Photopic luminance does not always predict perceived room brightness. Light. Res. Tech. 22(1):37-41. 131. Balder, J. J. 1957. Erwünschte Leuchtdichten in Büroräumen. Lichttechnik 9(9):455-461. 132. Bodmann, H. W. 1967. Quality of interior lighting based on luminance. Trans. Illum. Eng. Soc. (London) 32(1):22-40. 133. Bodmann, H. W. 1962. Illumination levels and visual performance. Int. Light. Rev. 13(2):41-47. 134. Bodmann, H. W., Sollner G., and E. Voit. [1964]. Bewertung von Beleuchtungsniveaus Bei Verschiedenen Licharten. Proceedings, Commission Internationale de l'Éclairage 15th Session, Paris: Bureau Central de la CIE. 135. Boyce, P. R. 1973. Age, illuminance, visual performance and preference. Light. Res. Tech. 5(3):125145. 136. Hughes, P. C., and J. F. McNelis. 1978. Lighting, productivity, and the work environment. Light. Des. Appl. 8(12): 32-40. 137. Nemecek, J., and E. Grandjean. 1973. Results of an ergonomic investigation of large-space offices. Hum. Factors 15(2):111-124.

138. Saunders, J. E. 1969. The role of the level and diversity of horizontal illumination in an appraisal of a simple office task. Light. Res. Tech. 1(1):37-46. 139. Bean, A. R., and A. G. Hopkins. 1980. Task and background lighting. Light. Res. Tech. 12(3):135-139. 140. Adrian, W., and K. Eberbach. 1969. On the relationship between the visual threshold and the size of the surrounding field. Light. Res. Tech. 1(4):251-258. 141. Bisele, R. L. Jr. 1950. Effect of task-to-surround luminance ratios on visual performance. Illum. Eng. 45(12):733-740. 142. Lighting and human performance: A review. 1989. Washington DC: National Electrical Manufacturers Association. 143. McCann, J.J., and J. A. Hall. 1980. Effect of average luminance surround on the visibility of sine wave gratings. J. Opt. Soc. Am. 70(2): 212- 219. 144. Wilson, A. J. and A. Lit. 1981. Effects of photopic annulus luminance level on reaction times and on the latency of evoked cortical potential responses to target flashes. J. Opt. Soc. Am. 71(12): 1481-1486. 145. Cobb, P. W., and F. K. Moss. 1928. The effect of dark surroundings upon vision. J. Franklin Inst. 206 (6):827-840. 146. Johnson, H. M. 1924. Speed, accuracy, and constancy of response to visual stimuli as related to the distribution of brightnesses over the visual field. J. Exp. Psychol. 7(1):144. 147. Luckiesh, M. 1944. Brightness engineering. Illum. Eng. 39(2):75-92. 148. Rea, M. S., M. J. Ouellette, and D. K. Tiller. 1990. The effects of luminous surroundings on visual performance, pupil size, and human preference. J. Illum. Eng. Soc. 19(2):45-58. 149. Slater, A. I., and P. R. Boyce. 1990. Luminance uniformity on desks: Where is the limit? Light. Res. Tech. 22(4):165-174. 150. Tuow, L. M. C. 1951. Preferred brightness ratio of task and its immediate surroundings. Proceedings: Commission Internationale de l'Éclairage 12th Session. Paris: Bureau Central de la CIE. 151. Tregenza, P. R., S. M. Romaya, S. P. Dawe, L. J. Heap, and B. Tuck. 1974. Consistency and variation in preferences for office lighting. Light. Res. Tech. 6(4):205-211. 152. van Ooyen, M. H. F., J. A. C. van de Weijgert, and S. H. A. Begemann. 1987. Preferred luminances in offices. J. Illum. Eng. Soc. 16(2):152-156. 153. Roll, K. F., and H. J. Hentschell. 1987. Luminance patterns in interiors and balanced perception. Proceedings: 21st session. Commission Internationale de l'Éclairage. Paris: Bureau Central de la CIE. 154. Kruithof, A. A. 1941. Tubular luminescence lamps for general illumination. Philips Tech. Rev. 6 (3):65-73. 155. Baron, R. A., M. S. Rea, and S. G. Daniels. 1992. Effects of indoor lighting (illuminance and spectral distribution) on the performance of cognitive tasks and interpersonal behaviors: The potential mediating role of positive affect. Motiv. Emot. 16(1):1-33. 156. Boyce, P. R., and C. Cuttle. 1990. Effect of correlated colour temperature on the perception of interiors

and colour discrimination performance. Light. Res. Tech. 22(1):19-36. 157. Davis, R. G., and D. N. Ginthner. 1990. Correlated color temperature, illuminance level, and the Kruithof curve. J. Illum. Eng. Soc. 19(1):27-38. 158. Kallman, W. M., and W. Isaac. 1977. Altering arousal in humans by varying ambient sensory conditions. Percept. Mot. Skills 44(1):19-22. 159. Delay, E. R., and M. A. Richardson. 1981. Time estimation in humans: Effects of ambient illumination and sex. Percept. Mot. Skills 53(3):747-750. 160. Wilkinson, R. 1969. Some factors influencing the effect of environmental stressors upon performance. Psychol. Bul. 72(4):260-272. 161. Taylor, L. H., and E. W. Socov. 1974. The movement of people toward lights. J. Illum. Eng. Soc. 3 (3):237-241. 162. Yorks, P., and D. Ginthner. 1987. Wall lighting placement: Effect on behavior in the work environment. Light. Des. Appl. 17(7):30-37. 163. LaGiusa, F. F., and L. R. Perney. 1974. Further studies on the effects of brightness variations on attention span in a learning environment. J. Illum. Eng. Soc. 3(3):249-252. 164. Sanders, M., J. Gustanski, and M. Lawton. 1974. Effect of ambient illumination on noise level of groups. J. Appl. Psychol. 59(4):527-528. 165. Kallman, W. M., and W. Isaac. 1977. Altering arousal in humans by varying ambient sensory conditions. Percept. Mot. Skills 44(1):19-22. 166. Badia, P., B. Meyers, M. Boecker, J. Culpepper, and J. R. Harsh. 1991. Bright light effects on body temperature, alertness, EEG and behavior. Physiol. Behav. 50(3):583-588. 167. Boyce, P. R., J. W. Beckstead, N. H. Eklund, R. W. Strobel, and M. S. Rea. 1997. Lighting the graveyard shift: The influence of a daylight-simulating skylight on the task performance and mood of nightshift workers. Light. Res. Tech. 29(3):105-134.

4 Color Architects, engineers, interior and industrial designers, colorists and color stylists, and lighting designers all need to understand color. This chapter has been prepared to increase mutual understanding among those responsible for creating the environment and making it visible and visually functional. Electromagnetic radiant energy provides a physical stimulus that enters the eye and causes the sensation of color (see Chapter 3, Vision and Perception). The spectral characteristics of the stimulus are integrated by the visual system and cannot be differentiated without the use of an instrument. Because the color and the color rendering properties of light sources are increasingly important in the design of an illuminated environment, lighting designers need a good working knowledge of the vocabulary and practices of modern color science. The aesthetic use of color to produce pleasing interiors requires coordination between the interior designer and the person designing the lighting. Each needs to know how to use color to help provide the desired brightness levels and distributions. Today's lighting designer is faced not only with a choice of color in light sources but also with wide variations in color rendering properties of light sources that can be identical in color.1 To provide lighting designers with a basis for their studies in color, the IESNA committees have developed several reports2-5 that provide useful background material for this chapter. In addition, the chapter concludes with examples of several fields of special applications. Other chapters contain brief discussions of color, with specialized applications. Information on colorimetry of light sources is not contained in this chapter, but is found in IES LM-16-1984.2 Color is a fundamental parameter of vision and perception. Discussions related to color threshold discrimination, color vision abnormalities, visual processing channels, and perceptions of lighting are provided in Chapter 3, Vision and Perception.

BASIC CONCEPTS OF COLOR Color Terms In the Glossary, color terms are defined carefully to provide a way of distinguishing between several commonly confused meanings of the word "color." Whether one makes strict use of the definitions or not, an understanding of the purpose and need for the differentiations that are made is basic to an understanding of the subject. For additional information on color, see Plates 1, 2, and 3 at the end of this chapter. The perceived color, the color perceived as belonging to an object or light source, is something perceived instantaneously. It is so common an experience that many people find it hard to understand why color is not simple to explain in a few easy lessons. But a color perception results from the complex interaction of many factors including the characteristics of the object or light source, the light incident on an object, the surround, the viewing direction, observer characteristics, and the observer's adaptation. Characteristics of object, light, surround, and observer can vary both spectrally and directionally, each in a different manner. The observer might vary in regard to time of seeing, what was seen last, or how attention was focused in relation to the time of seeing. Unless the circumstances of a former situation with which the layperson, interior designer, or lighting designer might be familiar are similar enough in all important respects, a new situation cannot be responded to by reference to past experience alone. Laypersons can cope with a new situation by making certain assumptions or by limiting themselves to the use of conditions with which they are familiar. But lighting designers cannot do this if they are to deal with all types of architectural situations, with all types of light sources, and with requirements that will fit new or specialized situations. Color (sometimes called psychophysical color) is defined as the characteristic of light by which an observer can distinguish between patches of light of the same size, shape, and structure. It reduces itself to a basic description of light in terms of amounts of radiant power at the different wavelengths of the visually effective spectrum, which for most practical purposes is considered to extend from 380 to 780 nm. (To identify colors due in part to fluorescent dyes activated by energy in the ultraviolet [UV] region, it is necessary in specifying the spectral distribution of a light source to extend the wavelength range beyond that which is visually effective, down to 300 nm in the UV region, particularly for sources that are intended to reproduce daylight.) Identical colors are produced not only by identical spectral power distributions (SPDs) but also by many different SPDs. Such different SPDs are called metamers. The color of an object, or object color, is defined as the color of light reflected or transmitted by an object when it is illuminated by a standard light source. For this purpose, a Commission Internationale de l'Éclairage (CIE) standard observer, using standardized conditions of observation, must be assumed. The word "color" often is used to cover all three meanings discussed above. When the assumed standard conditions are satisfied, then there is little need for distinguishing between the perceived color, the psychophysical color, and the object color. However, if designers are to handle new problems in color, including new light sources that can vary widely in spectral distributions, they must know the differences between the meanings of color and keep these distinctions in mind even when using the one term to cover all three. The term "color temperature" is widely used--and often misused--in illumination work. It relates to the color of a completely radiating (blackbody) source at a particular temperature and of light sources that color-match such a body. The color temperature of a light source is the absolute temperature of a blackbody radiator having a color equal to that of the light source. Its correlated color temperature is the absolute temperature of a blackbody whose color most nearly resembles that of the light source.

Abnormal Color Vision Approximately 8% of males and 0.4% of females have color vision that differs from that of the majority of the population. These people are usually called "color blind," although very few (less than 0.01% of the total population) can see no color at all.6 Most color-blind people can distinguish yellows from blues but confuse reds and greens. Their data should be excluded from any color measurements or color evaluation procedures that are to be used for application to the general population. See Chapter 3, Vision and Perception, and Figure 3-21 for additional discussion and data on abnormal color vision.

Color Rendering Color rendering is a general expression for the effect of a light source on the color appearance of objects in conscious or subconscious comparison with their color appearance under another (reference) light source. Methods of measuring and specifying color rendering properties of light sources depend on the color appearance of objects under a reference, or standard, light source compared with the appearance of the same objects under the test source. The color rendering properties of a light source cannot be assessed by visual inspection of the source or by a knowledge of its color.7 For this purpose, full knowledge of its SPD is required. Viewed in succession under lamps that look quite alike but are different in spectral distribution, objects might look entirely different in color. An extreme case is a pair of color-matched low-pressure sodium and yellow fluorescent lamps. Most objects, which in daylight might look red, yellow, green, blue, or purple, will appear quite different under these two lamps. Under the sodium lamp objects will lose their daylight appearance,

appearing more or less as one hue, from light to very dark (near-black). Under the yellow fluorescent lamp, more hues can be recognized, but the color of objects will still differ considerably from their daylight color.

Basis for Measurements Because color is the characteristic by which a human observer distinguishes patches of light, and light is visually evaluated radiant energy, color can be computed by combining physical measurements of radiant power, wavelength by wavelength, with data on how an observer matches colors. The color matching characteristics of the internationally adopted CIE standard observers, defined by the tristimulus values of an equal-power spectrum, are provided in Figures 4-1 and 4-2. These are the color matching functions. With data for a standard observer and the spectroradiometric measurement of a light source, the chromaticity of that light source can be calculated. Thus spectroradiometry becomes a tool for color measurement. Measurements of radiant power are physical, while evaluation of radiant power by a human observer, based solely on perception, is psychological. Visual evaluations, quantified through measurements made for standardized conditions of test, provide psychophysical methods of measurement. Visual evaluation of the appearance of objects and light sources can be in terms derived wholly from one's perceptions. One convenient and useful set of terms describing these perceptions for light sources is hue, brightness, and saturation.8 Hue is the attribute according to which an area appears to be similar to one, or to proportions of two, of the perceived colors red, yellow, green, and blue. Brightness is the attribute according to which an area appears to be emitting more or less light. Saturation is the attribute by which an area appears to exhibit more or less chromatic color (that is, departure from gray), judged in proportion to its brightness. Many widely used psychophysical methods for describing and specifying color show poor correlation with perceptual factors, and often these are converted to more meaningful visual terms, usually to a more uniform color spacing, of which the Munsell system9 and the CIE 1976 Uniform Color Spaces10,11 are prime examples.

CIE Method of Color Specification Basic CIE Method.10 This is a method originally recommended in 1931 by the CIE to define all metameric pairs by giving the amounts X, Y, Z of three imaginary primary colors required by a standard observer to match the color being specified. These amounts can be calculated as a summation of the spectral compositions of the radiant power of the source or the illuminated color specimen, times the spectral tristimulus values for an equal-power source (Figure 4-1). For example,

Figure 4-1. Color Matching Functions. (a) CIE 1931 Standard Observer (2°); (b) CIE 1964 Standard Observer (10°)

where S(λ) = spectral irradiance distribution of the source (Figure 4-3) ρ(λ) = spectral reflectance of the specimen, k = a normalizing factor

¯ x(λ) = spectral tristimulus value from Figure 4-1 with similar expressions for Y and Z, wherein ¯ y(λ) and ¯ z(λ) respectively are substituted for ¯ x(λ). The normalizing factor k can be assigned any arbitrary value provided it is kept constant throughout any particular application. Where only the relative values of X, Y, and Z are required, the value of k is usually chosen so that Y has the value 100.0. In the special case where the absolute values of S(λ) ∆λ are given (for example, in watts), it is convenient to take k = Km = 683 lm/W, whereby the value of Y gives the equivalent luminous quantity in lumens. Here, the accepted symbol for S(λ) is Φe,λ (see Glossary under luminous flux). For colors of reflecting objects, the reflectance factor, R(λ), must be introduced, so that

In this case, the normalizing factor k is usually given the value

Figure 4-2. Graph of CIE 1931 and 1964 color-matching functions. (y-axis represents the tristimulus values.) With this normalization, the value of Y is the luminous reflectance factor expressed in percent. Use of the reflectance factor, R(λ), is appropriate for calculating tristimulus values that relate to the appearance of objects. For such other applications as calculations of light flux in a space, the reflectance, ρ(λ), might be more appropriate. (For transmitting objects, the transmittance or transmittance factor must be used in place of the reflectance or reflectance factor.) Precise definitions of R(λ) and ρ(λ) are given in the Glossary under "reflectance of a surface or medium" and "reflectance factor." The essential difference between the two quantities is that the reflectance factor is directional and is measured relative to a perfect diffuser. Consequently it might be greater than 1 in certain directions as long as this is compensated by values less than 1 in other directions. The CIE has recommended two sets of standard color matching functions. The first is known as the CIE 1931 Standard Observer (Figure 4-1a) and is intended for use when the angular subtense of the field of view is between 1 and 4°. The second is the CIE 1964 Standard Observer (Figure 4-1b), intended for use with angular subtenses greater than 4°. The CIE has also recommended several standard illuminants (spectral power distributions) for use in computing object colors. The most commonly used are listed in Figure 4-3. They include Standard Illuminant A (representing a blackbody radiator at a color temperature of 2856 K; see also Figure 2-5), Standard Illuminant C (based on a laboratory simulation of average daylight), and Standard Illuminant D65 (a more modern and preferred representation of a phase of daylight at a correlated color temperature of approximately 6500 K). In addition, the CIE recommends a calculation method for standard illuminants representing phases of daylight at any correlated color temperature between 4000 and 25,000 K. In its recommendation, the CIE distinguishes between illuminants and sources. The term "source" refers to a physical emitter of light, such as a lamp or the sun and sky. The term "illuminant" refers to a specific SPD. The most accurate calculation method recommended by the CIE is summation at 1-nm intervals from 360 to 830 nm. However, the color matching functions have relatively small values at the ends of the spectrum, and furthermore, many sources and objects have fairly smooth spectral distributions so that summation from 380 to 780 nm at 5-nm intervals will suffice for many practical purposes, allowing the use of simpler instrumentation and computation. If data are available only for a restricted wavelength range (for example, 400 to 700 nm) or for a wider wavelength interval (for example, 10 or 20 nm), the appropriate values can be selected from Figures 4-1 and 4-3. An accurate method for dealing with incomplete data is to use special tables given in ASTM Standard Method E 308, Computing Colors of Objects by Using the CIE System.12 An important practical consideration for such sources as discharge lamps, which do not have smooth SPDs, is that the measurement bandwidth should be a multiple of the wavelength interval. The fractions X/(X + Y + Z), Y/(X + Y + Z), and Z/(X + Y + Z) are known as the chromaticity coordinates, x, y, z, respectively. Note that x + y + z = 1, and specification of any two fixes the third. By convention, chromaticity usually is stated in terms of x and y and plotted in a rectangular coordinate system as shown in Figure 4-4. In this chromaticity diagram, the points representing light of single wavelengths plot along a horseshoe-shaped curve called the spectrum locus. The line joining the extremities of the spectrum locus is known as the purple boundary and is the locus of the most saturated purples obtainable. A sample calculation for determining the CIE coordinates is shown in Figure 4-5 for a deep-red surface when illuminated by CIE illuminant D65. In Figure 4-5, column I is a listing of wavelengths in 5-nm steps, column II is a tabulation of spectral reflectance values for the deep-red surface at each wavelength in column I, and column III lists the CIE tristimulus computational data for CIE illuminant D65. By multiplying the row entry in column II by the corresponding one in column III and summing the products given in column IV, the values of X, Y, and Z are determined. Then by using the three fractions above, the chromaticity coordinates are determined. The percentage luminous reflectance is determined by multiplying the Y value by the normalizing factor k = 100/ΣS(λ)¯ y(λ)∆λ = 0.0095 A final recommendation of the CIE concerns geometrical arrangements for measuring colors of reflecting objects. Four alternative conditions for illuminating/viewing a test sample are specified: (1) 45°/normal, (2) normal/45°, (3) diffuse/normal, and (4) normal/diffuse (diffuse illuminating or viewing is

usually achieved by placing a sample in an integrating sphere). Consult Wyszecki and Stiles's Color Science13 for an extended discussion of the calculation and application of CIE data, including extensive tables of quantitative data and methods of colorimetry.

Figure 4-3. Spectral Power Distributions of CIE Standard Illuminants

Figure 4-4. The 1931 CIE chromaticity diagram showing method of obtaining dominant wavelength and purity for different samples under different light sources. Dominant Wavelength and Excitation Purity. Dominant wavelength and excitation purity are quantities more suggestive of the color appearance of objects than a CIE x, y specification and can be determined on an x, y diagram in relation to the spectrum locus and an assumed achromatic point (for object colors this is usually the point for the light source) (Figure 4-4). The dominant wavelength of all colors whose x, y coordinates fall on a straight line connecting the achromatic point with a point on the spectrum locus is the wavelength indicated at the intersection of that line with the spectrum locus. For some colors, the straight line from the achromatic point through the test chromaticity will strike the purple boundary rather than the spectrum locus. For these colors the line must be extended backwards from the achromatic point. The point where the extended line strikes the spectrum locus determines the complementary wavelength of such a color. The excitation purity is defined as the distance from the achromatic point to (x, y) divided by the total distance in the same direction from the achromatic point to the spectrum locus or the purple boundary.

Figure 4-5. Continued

Figure 4-5. Determination of CIE Chromaticity Coordinates from the Spectrophotometric Curve for a Surface Illuminated by Standard Illuminant D65 An x, y specification of any object color relates it only to the light source for which the object color is calculated. Consequently, the dominant wavelength and excitation purity of any object depend on the spectral composition of its illumination. CIE Uniform Color Spaces. Distances in the CIE x, y diagram or X, Y, Z space do not correlate well with the perceived magnitudes of color differences. This gives rise to the shapes and different sizes of MacAdam ellipses (Figure 3-35) which set the bounds for threshold discrimination between two colors; one at the center of the ellipse and the other anywhere on the edge of the ellipse. In a uniform color space, the ellipses would appear as circles of equal radii. Various transformations have been suggested that provide more uniform spacing. In 1960, the CIE provisionally recommended that whenever a diagram is desired to yield chromaticity spacing more uniform than the CIE x, y diagram, a uniform chromaticity-scale diagram (CIE 1960 UCS Diagram) based on that described in 1937 by MacAdam14 be used. The ordinate and abscissa of this u, v diagram are defined as

In 1976, after additional investigation, the CIE modified its 1960 UCS Diagram as follows:

Figure 4-6 illustrates the CIE 1976 UCS Diagram. To convert the CIE 1960 UCS Diagram to a three-dimensional system that is useful in studying color differences, the CIE, in 1964, added a recommendation developed for the purpose by Wyszecki15 that converts Y to a lightness index, W*, by the relationship

and converts the chromaticity coordinates u, v to chromaticness indices U, V by the relationships

The lightness index W* approximates the Munsell value function in the range of Y from 1 to 100%. The chromaticity coordinates un, vn refer to the nominally achromatic (neutral) color (usually that of the source) placed at the origin of the U*, V* system. In 1976, the CIE10,11 recommended two new uniform color spaces, known as CIELUV and CIELAB. Although these give a more uniform representation of color differences and therefore supersede the U*, V*, W* space for most purposes, the earlier system is still used for the calculation of CIE color rendering indices. Two spaces were recommended, rather than one, because experimental evidence was insufficient to select a single space that would be satisfactory for most industrial applications.

Figure 4-6. The CIE 1976 UCS diagram. The three coordinates of CIELUV are L*, u*, and v*, defined by

where u'= 4X/(X + 15Y + 3Z), v'= 9Y/(X + 15Y + 3Z), u'n, v'n, Yn = values of u', v', and Y for the nominally achromatic color (usually that of the source with Yn = 100). The major change from the U*, V*, W* system is that v'= 1.5v. The quality L* is a minor modification of W*; u' is the same as u. The three coordinates of CIELAB are L*, a*, and b*, defined by

where

(with q = X/Xn, Y/Yn, or Z/Zn). Here Xn, Yn, and Zn are the values of X, Y, and Z for the nominally achromatic color (usually that of the source with Yn = 100). The lightness index L* is the same for both CIELUV and CIELAB. Loci of constant Munsell hue and chroma for value 5/ (see discussion of Munsell Color System below) are plotted in u*, v* and a*, b* diagrams16 in Figure 4-7. The Munsell Color System is often used to test if the color space associated with a color-difference formula provides uniform spacing. If either diagram provided uniform spacing of the Munsell system, these loci would be straight, equally spaced radial lines and concentric, equally spaced circles. While neither diagram is perfect in this respect, Robertson noted that the Munsell data represent color differences much larger than threshold and are not necessarily suitable for comparing color-difference formulas that are intended to quantify near-threshold differences. He concluded that insufficient data are available to determine which of the color difference formulas is best.11 These two uniform color spaces each have associated with them a color-difference formula by which a measure of the total difference between two object colors can be calculated. In the CIELUV system, the color difference is measured by

In the CIELAB system it is measured by

These two formulas are useful for setting color tolerances in industrial situations. They are recommended by the CIE to unify practice, which in the past has involved the use of 10 or 20 different color-difference formulas. Correlates of the subjective attributes lightness, perceived chroma, and hue can be derived from either CIELUV or CIELAB as follows:

Although these quantities are approximate correlates of the respective subjective attributes, the actual perceived color depends significantly on the viewing conditions, for example, the nature of the surround. The exact degree of agreement of these measures with the corresponding subjective attributes, even for standard daylight viewing conditions, has not been determined. In commercial situations involving small color differences, tolerances often are set differently for L*, C*, and h because the acceptability can be different for the three components. The geometrical relationships among CIELAB coordinates are illustrated in Figure 4-8. The relationships in CIELUV (Figure 4-9) are similar. The 1976 CIELAB-CIELUV recommendation has been much more successful than the 1964 (U*, V*, W*) convention. Both formulas are in widespread use, the choice between them being based mainly on practical considerations other than uniformity of spacing. In industries concerned with such self-luminous colors as television screens and video displays, both CIELUV (Figure 4-8) and CIELAB (Figure 4-9) have been used. The same has been true of industries concerned with object colors. Indeed, some formulas (e.g., the CMC formula17,18 developed since 1976) continue to use CIELAB as a base and add extra complexity to improve the fit to visual acceptability data.

Figure 4-7. Loci of constant Munsell hue and chroma plotted in the CIE 1976 u*v* diagram (a), and the CIE 1976 a*b* diagram (b).

Other Systems of Color Specification Munsell System. This is a system of specifying color on scales of hue, value, and chroma. The hue scale consists of 100 steps in a circle containing five principal and five intermediate hues. The value scale contains ten steps from black to white, 0 to 10. The chroma scale can contain 20 or more steps from neutral gray to highly saturated. Each of the three scales is intended to represent equal visual intervals for a normal observer fully adapted to daylight viewing conditions (CIE source C) with gray to white surroundings. Under these conditions the Munsell hue, value, and chroma of a color correlate closely with the hue, lightness, and perceived chroma of color perception; under other conditions the correlation is lost. It is only for daylight conditions that Munsell samples are expected to appear equally spaced. When problems of color adaptation are fully solved, it might be possible to calculate the change in appearance and spacing that takes place when samples are viewed under a light source of different SPD.

Figure 4-8. Sketch of CIE 1976 (L*a*b*) color space with outer boundary generated by optimal color stimuli with respect to CIE standard illuminant D65 and the CIE 1964 supplementary standard observer. The colors of all object-color stimuli fall within this boundary. This is also the gamut within which the CIE 1976 color-difference formula ∆E(L*a*b*) is intended to be valid. Note that the spectrum locus of the monochromatic stimuli is generally well outside the boundary of object-color stimuli. From G. Wyszecki and W. Stiles, Color science. Copyright © 1982. Reprinted by permission of John Wiley & Sons, Inc.

Figure 4-9. Sketch of CIE 1976 (L*u*v*) color space with outer boundary generated by optimal color stimuli with respect to CIE standard illuminant D65 and the CIE 1964 supplementary standard observer. The colors of all object-color stimuli fall within this boundary. This is also the gamut within which the CIE 1976 color-difference formula ∆E(L*u*v*) is intended to be valid. Note that the spectrum locus of monochromatic stimuli is generally well outside the boundary of object-color stimuli. From G. Wyszecki and W. Stiles, Color science. Copyright © 1982. Reprinted by permission of John Wiley & Sons, Inc. Munsell notation is useful whether or not reference is made to Munsell samples. It has the form [hue] [value]/[chroma], for example, 5R 4/10. This is read "5 red, 4 over 10" or "5 red, 4 slash 10." Colors of zero chroma, which are known as neutral colors, are written N1/, N2/, etc., as shown in Figure 4-10. One widely used approximation of equivalence between hue, value, and chroma units is 1 value step = 2 chroma steps = 3 hue steps (when the hue is at chroma 5).

Figure 4-10. Cut-away view of the Munsell color solid showing notation scales of hue, value, and chroma (for example, 5Y 5/4), and the relation of constant hue charts to the three-dimensional representation. The Munsell scales are exemplified by a collection of color chips forming an atlas of charts that show linear series for which two of the three variables are constant (Figure 4-10). For use as standards or in technical color control, collections of carefully standardized color chips in matte or glossy surface can be obtained from Munsell Color Company, c/o Macbeth, P.O. Box 230, Newburgh, NY 12550, in several different forms. Since 1943 the smoothed renotation for the system, recommended by the Optical Society of America's Colorimetry Committee, has been recognized as the primary standard for these papers.

Instructions for obtaining Munsell values by calculation, or by conversion through CIE are contained in several publications.6,9,19 The relationship between Munsell value and CIE luminous reflectance factor is summarized in Figure 4-11.

Figure 4-11. Relationship between Munsell Value and Luminous Reflectance Factor ISCC-NBS Method of Designating Colors. The Inter-Society Color Council-National Bureau of Standards method of designating colors appeared in its original form in 1939 as NBS Research Paper RP 1239. The second edition appeared in book form in 1955 as NBS Circular 553, usually called the Color Names Dictionary (CND). The first Supplement to the CND, called the Centroid Color Charts (1965),20 provides useful low-cost color charts that illustrate, with 1-in.2 samples, the centroid color for as many (251) of the 267 color names in the system as could be matched at that time. Each of the names defines a block in color space. This method is distinguished from all others in that the boundaries of each name are given, rather than points. These boundaries are defined in Munsell notation. A method for pinpointing colors is not provided, but the system does give an understandable color description. When close distinctions must be made between samples that might bear the same ISCC-NBS designation, such specifications as CIE or Munsell should be used. The method is simple in principle: terms "light," "medium," and "dark" designate decreasing degrees of lightness, and the adverb "very" extends the scale to "very light" and "very dark"; adjectives "grayish," "moderate," "strong," and "vivid" designate increasing degrees of saturation. These and a series of hue names, used in both noun and adjective forms, are combined to form names for describing color in terms of its three perceptual attributes: hue, lightness, and perceived chroma. A few adjectives are added to cover combinations of lightness and perceived chroma: "brilliant" for light and strong, "pale" for light and grayish, and "deep" for dark and strong. The hue names and modifiers are listed in Figure 4-12. The second supplement to the CND, entitled The Universal Color Language (UCL), was published also in 1965.21 The UCL serves as the means of updating the CND. It brings together all the well-known color-order systems and methods of designating colors and interrelates them in six correlated levels of fineness of color designation, each higher level indicating a finer division of the color solid. It follows closely and extends the original requirements of the ISCC-NBS method of designating colors in the CND. The CND and the UCL have been published together as NBS Special Publication SP 440, with the UCL illustrated in color.22

Figure 4-12. ISCC-NBS Standard Hue Names and Modifiers OSA UCS System. The Optical Society of America (OSA) has produced a set of 558 color chips to illustrate uniform visual spacing on a regular rhombohedral lattice.23 Each chip is intended to be equally different from its 12 nearest neighbors in the lattice. Because of the noneuclidean nature of color space, perfectly uniform spacing is impossible to achieve in a three-dimensional lattice. Thus, the OSA Committee on Uniform Color Scales was forced to make some compromises in specifying the colors. These compromises are not evident on casual study, although they can be seen in more careful analyses. The set is sold by the Optical Society of America, 2010 Massachusetts Avenue NW, Washington, DC 20036, and has generated much interest, especially among artists, designers, and color scientists. Natural Color System (NCS). The NCS24 is based on a principle entirely different from that of the Munsell or the OSA system. The principle is that of resemblances to six elementary perceived colors: red, yellow, blue, green, black, and white. Of these, the four chromatic colors are those in which no trace of

the others can be seen. In a geometric representation, they are placed 90° apart on a hue circle. Black and white are perceived colors that contain no trace of each other or of any of the four chromatic colors. They are placed at the apexes of two opposite cones with their bases on the hue circle. Any color resembles at most two of the chromatic elementary colors plus black and white. It is claimed that the degree of resemblance can be estimated to within approximately 5%, even by naive observers. DIN System. The DIN color system25 is the official German standard. It is organized in terms of hue (Färbton), saturation (Sattigung), and darkness (Dunkelstufe). The system is defined in terms of CIE chromaticity and luminance factors with certain compromises made to keep the relationships as simple as possible. It attempts to show uniform steps of color difference and uses CIE colorimetry extensively for interpolation and extrapolation. Correlation Among Methods. Frequently it is desirable to convert from one system of specification to another, or to convert or identify the color of samples on a chart or color card to terms of another. If the coordinates or samples of one system are given in CIE or Munsell terms, they can be converted or compared to any other system for which a similar conversion is available. Color charts of the German standard 6164 DIN system are provided with both CIE and Munsell equivalents. The Japanese standard system of color specification, JIS Z 8721-1958, is in terms of hue, value, and chroma of the Munsell renotation system, according to the CIE x, y coordinates recommended by the Optical Society of America's 1943 subcommittee report.9 The name blocks of the ISCC-NBS method are in terms of the Munsell renotation system with samples measured in CIE terms. Having a common conversion language helps promote international cooperation and understanding of the subject. Complete sets of CIE-Munsell conversion charts are contained in ASTM Test Method D1535.19 Many of the available conversions are referenced in a 1957 paper by Nickerson.26 For more detailed descriptions of color systems or conversions, consult Color in Business, Science and Industry.6 For a useful survey of color order systems, consult Reference 27.

Color Temperature Blackbody characteristics at different temperatures are defined by Planck's radiation law (see Chapter 1, Light and Optics). The perceived colors of blackbody radiators at different temperatures depend on the state of adaptation of the observer. Plate 4 gives an approximate illustration of the perceived colors at various color temperatures for various states of adaptation; it shows that, as the temperature rises, the color changes from red to orange to yellow to white to blue.

Figure 4-13. CIE 1931 (x, y) chromaticity diagram showing lines of constant correlated color temperature, in kelvin, together with three standard illuminants: A, C, and D65). The locus of blackbody chromaticities on the x, y diagram is known as the planckian locus. Any chromaticity represented by a point on this locus can be specified by color temperature. Strictly speaking, color temperature should not be used to specify a chromaticity that does not lie on the planckian locus. However, what is called the correlated color temperature (the temperature of the blackbody whose chromaticity most resembles that of the light source) is sometimes of interest. The correlated color temperature can be determined from diagrams28 similar to the one shown in Figure 4-13, either by graphical interpolation or by a computer program.29 It should be noted that the concept becomes less meaningful as the distance from the planckian locus increases. Equal color differences on the planckian locus are more nearly expressed by equal steps of reciprocal color temperature than by equal steps of color temperature itself. The usual unit is the reciprocal megakelvin (MK−1), so that the reciprocal color temperature is 106 divided by the color temperature in kelvin (K). The term "mired" (pronounced mi' red), an abbreviation for "micro-reciprocal-degree," was formerly used for the unit. A difference of 1 reciprocal megakelvin indicates approximately the same color difference anywhere on the color temperature scale above 1800 K; yet it corresponds to a temperature difference that varies from approximately 4 K at 2000 K to 100 K at 10,000 K. Color temperature is a specification of chromaticity only. It does not represent the SPD of a light source. Chromaticities of many "daylight" lamps plot very close to the planckian locus, and their colors can be specified in terms of correlated color temperature. However, this specification gives no information about SPD, which can, and often does, depart widely from that of daylight. In particular, the addition of light from two sources each having blackbody distribution but different color temperatures does not produce a blackbody mixture. Figure 4-14 shows spectral curves for planckian distributions for different color temperatures. Distributions based on daylight30 are also available for several correlated color temperatures (see Figure 8-1 in Chapter 8, Daylighting). Most tungsten filament lamps approach the relative SPD of a blackbody quite closely. The color temperature of such lamps varies with the current passing through them. By varying the voltage across such a lamp, a series of color temperatures can be obtained covering a wide range up to approximately 3600 K.

Color Constancy and Adaptation4,31,32 A nonluminous colored object contributes to observed color by modifying the SPD of the light radiated to the eye. The color of the light reflected or transmitted by the object when it is illuminated by a standard light source is known as the object color and can be calculated by assuming certain conventions (as in the CIE system). The color seen when the object is viewed normally in daylight is a perceptual phenomenon referred to as the perceived color of the object. While there are exceptions, the perceived colors of objects, when illuminated by various sources, do not change as much as might be expected from the calculated difference

in chromaticities. This phenomenon is known as color constancy. Objects whose perceived colors change greatly when there is a wide change in illumination, as for example from daylight to incandescent filament light, are said to have unstable colors. It is important to remember that whereas the perceived color of an object might not change much with a change of light source color, the object color (as specified for example by CIE chromaticity coordinates) will change. For example, a piece of white paper will appear white under both incandescent light and daylight, but the object color will be quite different in the two cases because the paper, being spectrally neutral, will have almost the same chromaticity as the source in each case.

Figure 4-14. Family of planckian distribution curves. The impression that the perceived colors of most objects do not change greatly with the SPD of the light source is due primarily to a low degree of spectral selectivity in daylight and incandescent sources. Color constancy is affected by such factors as awareness of the illuminant, persistence of memory of colors, consistency of attitude toward the object, and adaptation of the visual mechanism. Adaptation is, in effect, a rebalancing of the color response of the visual system as the spectral composition of the visual scene changes. Thus, adaptation tends to counteract the shift in chromaticity of the source and thereby preserves the appearance of object colors. However, there are cases where even slight residual shifts can be noticeable, annoying, or even intolerable. Such cases might be encountered with foodstuffs, with displayed merchandise, or in the grading of various commercial products. The facts of color constancy and adaptation are not yet known well enough to make possible the computation of color rendering properties of a lamp with sufficient accuracy except when the reference or standard lamp is required to have the same correlated color temperature as the test lamp. When it becomes possible to compute the effects of constancy and adaptation so that the results agree with the subjective experience, then it will be possible to calculate the color rendering properties of a lamp irrespective of its SPD. In the meantime, as will be seen later, the CIE color rendering index does make an allowance for chromatic adaptation, even though the allowance is not perfectly accurate. See the section "Perceptual Constancies" in Chapter 3, "Vision and Perception," for details on other types of perceptual constancy.

Color Appearance Models The CIE has worked for many years to develop a mathematical model for color appearance, especially for comparisons between different media such as CRT displays and color printers. Several models have been proposed, including one by Hunt33 and another by Nayatani et al.34 In 1997, the CIE recommended a model called CIECAM9735 which includes features from many of the common models.

Color Contrast Color contrast is sometimes used colloquially to describe the property by which two adjacent fields of equal luminance but different chromaticity can be distinguished from one another. It should be noted that color contrast is not a quantifiable parameter as is luminance contrast (Equations 3-6 to 3-8). The color separation between two fields is more correctly specified in terms of the CIELUV (Equation 4-15) or CIELAB (Equation 4-16) color difference formulas. Color appearance is affected markedly by the color of adjacent areas, particularly if one surrounds the other. For example, a color patch appears brighter (less gray) if it is surrounded by a large dark area. It appears dimmer (more gray) if it is surrounded by a similar light area. Juxtaposed areas also induce shifts in hue and saturation in one another. Hues shift in opposite directions in color space, tending to induce complementary hues. Similarly, saturation interacts, magnifying saturation differences in juxtaposed palates of color. In general, there tends to be a simultaneous and complex shift in all three attributes when colors are placed side by side.

Metameric and Conditional Color Matches If two lights are visually indistinguishable because they have the same spectral compositions, they are said to form a spectral match. However, two lights can be visually indistinguishable in spite of having quite different SPD. Such a color match is said to be metameric, and the lights to be metameres. In the CIE system, the computed match is identified by application of color matching functions, which show that the tristimulus values for one light are identical to those for the other. If the lights are viewed by an observer characterized by different color matching functions, they might no longer match. All metameric matches are therefore conditional matches. The metameric character of a match sometimes will be revealed by looking at a spectrally selective object and noting that the object is of different color when illuminated by the two lights. This is illustrated in Figure 4-15. Objects with identical spectral reflectance distributions (see samples A and B in Figure 4-16) are said to produce an unconditional match. They match to everyone, no matter what source illuminates it. If, however, the color-matched reflected light comes from identically illuminated objects that have different

spectral reflectances (see samples E and F in Figure 4-16), the match is metameric. Substitution of another light source, or another observer, might upset the match; thus objects that can produce a metameric match, though identically illuminated, can be said to produce a match that is both observer conditional and source conditional. Such a match is illustrated in Plate 5.

Figure 4-15. Illustration of effect of metamerism of light sources. The two light sources, 1 and 2 in the bottom figure, have different spectral power distributions but are, themselves, metamers when illuminating a spectrally flat object. An object, such as the one whose spectral reflectance curve is shown in the top figure, may have different color appearances when illuminated by each of the two sources. Sometimes it has been argued wrongly that, because the presence of metamerism always corresponds to a conditional match, a metameric match is the same as a conditional match. Not all conditional matches, however, are metameric. Figure 4-16 shows the reflectance curves of two samples, C and D, that have different colors if the light source contains a significant amount of radiant power between 380 and 480 nm, but produce a match if the power of the source is confined to wavelengths greater than 500 nm. Samples C and D thus form a source-conditional match; no metamerism is involved because there is no source for which the SPD reflected from C and D have the same color but are spectrally different. There is a necessary though not sufficient condition that must be satisfied by the spectral reflectances of objects that, identically illuminated, can produce metameric reflected lights. First, the two reflectance curves must be different in some part of the visible spectrum, or else the reflected lights will be a spectral rather than a metameric match. Second, to be a color match the two objects must reflect equal amounts of the incident light, and this means that the curves must cross at least once within the visible spectrum. Third, the two objects must not differ in the yellow-blue sense (if the curves cross at only one wavelength within the visible spectrum, a yellow-blue difference is implied; therefore, the curves must cross at least at two wavelengths). Finally, the two objects must not differ in the purple-green sense (if the curves cross at only two wavelengths, a purple-green difference is implied). Therefore, the reflectance curves of objects capable of producing a metameric match for some combination of trichromatic observer and source must cross at least three wavelengths in the visible spectrum. Samples E and F in Figure 4-16 have this property and thus for some source-observer combination might produce a metameric pair of reflected lights, that is, they might match. For a discussion of the exact conditions under which a certain number of intersections is required, see Stiles and Wyszecki.36

Figure 4-16. Color matches. A and B are nonmetameric matches and will match for any observer under any light source. C and D will match for any observer under a source with no power at wavelengths less than 500 nm, but will not match for some observers under a source that does have some power in the wavelength region from 380 to 480 nm. C and D produce a source-conditional match, but they do not form a metameric pair, any more than A and B do, because the matching beams do not have different spectral compositions. E and F may form a metameric pair for some source-observer combinations.

USE OF COLOR Reflectance Every object reflects some fraction of the light incident upon it. The larger the fraction reflected, the "lighter" is the color of the object and the higher is the assigned Munsell value. In the Munsell color solid of Figure 4-10, the lightness dimension is in the vertical direction, along the scale of Munsell value, and ranges from black at the bottom to perfectly reflecting at the top. In the two-dimensional CIE chromaticity diagram of Figure 4-4 are plotted the two CIE coordinates that are related to Munsell perceived hue and to Munsell perceived chroma or saturation, constituting a horizontal plane in the color solid of Figure 4-10. Figure 4-17 shows how the lightness dimension relates to the chromaticity diagram. In the CIE system, the Y tristimulus value of the light reflected from an

object represents the luminous reflectance factor expressed as the percentage of the light that would be reflected by a perfectly reflecting diffuser. Computation of the percentage luminous reflectance is demonstrated in Figure 4-5, and its relation to Munsell value is shown in Figure 4-11.

Figure 4-17. The chromaticity diagram can be extended by adding a third axis for luminance factor. Lighter colors then appear directly above the points representing their chromaticity at a height representing their lightness. The luminous reflectance scale is not visually uniform between 0 and 100%, black and white. An object that reflects 50% does not look halfway between black and white, but looks much nearer to white. On the other hand, the purpose of the Munsell value scale is to illustrate equal visual steps for a given set of standard conditions. In Figure 4-11 (a condensed table), reflectance and Munsell value units are related. Thus, under daylight conditions, and for a light gray surround, a Munsell value of 5 should look approximately halfway in lightness between the black and white endpoints of the scale. Yet the luminance factor of a value-5 sample is only approximately 20%. A color with Munsell value 7 is called a light color, yet it reflects less than half (only 42%) of the light it receives. This is an important point for lighting designers to consider, for unless all the colors in the color scheme of a room layout are very light, well over 50% of the light is absorbed. If value-5 colors are used, as much as 80% of the incident light might be absorbed. With practice in the use of a Munsell value scale, particularly the special set of Munsell scales developed for lighting and interior designers, one can learn to estimate Munsell values rather accurately and convert them to luminous reflectance by means of Figure 4-11. Value-reflectance conversion tables for every Munsell value in steps of 0.1 are available in several publications6,9,19 or can be calculated as follows:6

where V = Munsell value, Y = luminous reflectance relative to a perfect diffuser (in %). Multiply Y by 1.0257 to convert it to the formerly used scale on which smoked magnesium oxide had the value of 100. The luminous reflectance of spectrally nonselective white, gray, and black objects remains constant for all light sources, but the luminous reflectances of colored objects will differ in accordance with the SPD of the light source. For example, with illumination from incandescent sources, which have relatively high radiant power in the middle- and long-wavelength portions of the visible spectrum and low power at the short-wavelength end, yellow objects appear lighter and blue objects darker than they do under daylight illumination; under blue sky the reverse will be true. On many sets of Munsell scales for judging reflectance, reflectances of each sample are given for three light sources: CIE A at 2856 K, CIE D65 at 6500 K, and cool white fluorescent at 4300 K. Light walls and ceilings, whether neutral or chromatic, are much more efficient than dark walls in distributing light uniformly. Step-by-step changes studied by Brainerd and Massey in 194237 have been reported in terms of illuminance and coefficients of utilization and are shown in Figure 4-18. Mathematical analyses by Moon38 on the effect of wall colors on illuminance and luminance ratios in cubical rooms show that an increase of wall reflectance by a factor of 9 can result

in an increase in illuminance by a factor of approximately 3. Moon38 has also published much information concerning spectral and colorimetric characteristics of materials used in room interiors.

Figure 4-18. Variation of illuminance and utilization coefficient with color scheme. The luminaire used for these results had a general diffuse distribution.

Figure 4-19. Graph showing how the spectral power distribution of sunlight when interreflected on a surface (spectral reflectance shown by the dashed line) changes with successive reflections in distribution and overall intensity. The top curve is the original source and the bottom the light after five reflections. When neutral and chromatic surfaces of equal reflectance are uniformly and directly illuminated, they will have equal luminances. But by interreflections in a room, the light reaching a working surface will have undergone several reflections from ceiling and walls, and the perceived colors of ceiling and walls, as well the light reaching the working surface, will have become more saturated. Figure 4-19 shows the measured spectral reflectance (dashed curve) of a pale pink surface. The SPD of incoming sunlight is represented by the topmost solid curve. The lower curves are the result, respectively, of one to five reflections from the pale pink surface. The remaining light deepens in color progressively. Figure 4-20 lists for the six SPDs the computed chromaticity, remaining power content, remaining lumen content, and perceived color of the light. It is clear that multiple reflections are costly in lumen content of the illumination and can cause unpleasant intensification of its color.

Figure 4-20. Characteristics of Illumination as It Enters a Space and After Successive Reflections Spencer and Sanborn,39 O'Brien,40 and Jones and Jones41 have published basic studies in this field. Spencer and Sanborn have analytically found the color shift due to interreflection in an infinite room and in a finite rectangular room, and O'Brien has developed and used computer methods and results to provide charts and tables to aid designers in making detailed predictions of illuminance and luminance distributions in rooms, a prerequisite for solving the problem for color interreflections. In France, Barthès has published experimental measurements for a model room.42 In Japan, Krossawa43 has computed data on a closed surface painted with a uniform color and derived a general empirical formula for the color shift due to interreflections for different colors. Yamanaka and Nayatani44 have compared results for computed and actual rooms, and consider the agreements to be quite satisfactory under the model conditions. Gradually the data based on such studies will reach a form in which the practicing designer can use them. Meanwhile designers should understand the general principles so that color change by room interreflections can be taken into consideration in planning a lighting layout.

Color Schemes: Choosing Suitable Colors No set of simple rules can allow for tastes of different people, or for different conditions and changing fashions. However, the following suggestions provide a place to start:

1. Ceilings are assumed to be white, or slightly tinted. (Note: Some hospital ceilings can be treated as a fifth wall for the supine patient.) 2. Walls, floors, and other structural elements, which will be changed infrequently, must be considered first in the color scheme. 3. Smaller areas (e.g., machinery or furniture) need only blend or contrast with walls and floors. 4. Color schemes, represented by material, surfacing, or paint (coating) samples, should be assembled and evaluated under lighting conditions closely duplicating those under which the scheme will be used. This will help avoid the problem of significant color shifts and of failures in metameric matches. 5. Because major surfaces can contribute considerably to the distribution of light by reflection and interreflection, the luminous reflectance (Munsell value) should be high where high task luminances are important. 6. The prime purpose of the color scheme needs consideration. Visibility can be most important in a schoolroom, dignity in a church, a sense of well being in a factory, an atmosphere of excitement in a circus, and quiet in an office. 7. Limitations can exist for redecoration schemes that must be built around existing colors of carpeting or other flooring, draperies, or furniture. The 1962 report of the Color Committee3 has three useful color charts. The first provides scales of hue, value, and chroma to help in understanding color terminology used in interior design. The second provides a series of 66 color chips arranged to show strong versus weak chromas and warm versus cool colors, with reflectances and Munsell notation for each sample, for colors used for interior surfaces. The third shows a 10-sample hue circle of typical wall colors at 60% reflectance, and three sample color scheme selections. To assist the designer, the following narrative describes current thinking on color schemes. Only limited research data support these conclusions, however. Color schemes usually are variations of basic plans classified as monochromatic (single-hued), complementary, adjacent or analogous, split complementary, or triads. The dominant character usually is determined by the largest area, and in three-hued schemes this usually is the least saturated. A large pattern, strong in value contrast, makes a room seem smaller; a small pattern, in gentle contrast and high reflectance, can make it seem larger. The absence of pattern can provide the illusion of maximum space. The effects of strong contrasts of color or pattern are similar to each other, that is, they both are stimulating, make people restless, and make time seem longer. They are effective for corridors, places of entertainment, entrance halls, public washrooms, quick lunch counters, and other locations where it is desired that people spend a short time; but they are undesirable in hospitals, for example. Gentle contrast is restful and makes time seem shorter. The play of molded form and texture can add interest; contrasts of natural wood, brick, stone, and woven materials add interest to smooth painted walls. Although personal tastes in color vary with climate, nationality, age, gender, and personality, there is almost universal agreement to call yellows, yellow-reds, reds, and red-purples warm colors, and to call greens, blue-greens, blues, and purple-blues cool colors. All grays approach neutral. The apparent size and position of objects are affected by color. High-chroma, warm colors usually are the most advancing, and cool colors the most receding. Lowering the chroma reduces the effect on apparent position. Light colors make objects appear larger; dark colors make objects appear smaller.

Selection Guide By considering such factors as warmth, spaciousness, and excitement level, it is possible to determine a suitable dominant color and degree of contrast. These considerations can be dealt with in four steps3 to help decide on values, hues, chromas, and contrasts in the Munsell system. Step 1: Value determination. This step helps decide the value, that is, how light or dark a color scheme should be. If a high level is necessary, colors with a high reflectance should be used. Dark colors tend to reduce luminance and contrast and produce luminance ratios that are unsatisfactory for efficient seeing. For areas in which illuminances of 750 lx (75 fc) or higher are recommended, the dominant values should be kept high, with reflectances of 40 to 60% or higher where the task is critical. Where lower illuminances are recommended, around 300 lx (30 fc), lower values can be introduced, at reflectances of 35 to 60%. For still lower levels, the dominant values can be even lower, with reflectances for large areas down to 15 to 35%. Where the best feasible visibility at low illuminances must be the goal, as for example in a parking garage, light colored (high-value) walls and ceilings are recommended. Step 2: Hue determination. Use warm, exciting, advancing colors where rooms have northern exposure, cool temperatures, and low noise element; where the room is too large and has smooth textures; where there is light physical exertion, time exposure is short, and a stimulating atmosphere is required; and where lamps are cool fluorescent. Use cool, restful, receding colors for rooms with southern exposure, warm temperatures and high noise element; for small rooms with rough texture; where physical exertion is heavy, time exposure is long, and a restful atmosphere is desired; and where lamps are incandescent or warm fluorescent. Step 3: Chroma determination. Strong chromas are used primarily for advertising, display, accents, and food merchandising; achromatic colors are primarily used for fashion areas, general interiors, and other merchandising. Use strong chromas if the time of occupant exposure is short, the general level of responsibility is low, a lively atmosphere is desired, the noise level is low, and a sense of taste or smell is unimportant. Use gray, low-chroma colors if the time exposure is long, the level of responsibility is high, an atmosphere of dignity is desired, the noise level is high, or a sense of taste or smell is important. Step 4: Contrast determination. Contrast is obtained by using light with dark, hues that are complementary, and low with high chromas. Little or no contrast should be used if the time of occupant exposure is long, the room size is small, a dignified atmosphere is required, or the wall surfaces are textured. Strong contrast should be used if the time of occupant exposure is short, the room size is large, a lively or exciting atmosphere is desired, or the wall surfaces are flat. These recommendations represent a consensus from working knowledge of designers and architects, based on field experience. Inasmuch as results from scientific investigations do not contradict them, they remain generally accepted. The designer must realize that in some cases a new or different approach can overrule common practice for a number of reasons.

Color Preference Research by Helson45-48 and others49,50 have added to our understanding of color preference in lighting. In his reports, Helson states the pleasantness of object colors depends on the interaction of the light source with the background color and with the hue, lightness, and saturation of the object color. The best background colors for enhancing the pleasantness of object colors were found to have either very high Munsell values (8/ or 9/) or very low values (2/ or 1/), and, with only one exception, very low or zero chroma. The background color was found to be more important than the SPD of the light source. Neutrals rank high as background colors, but very low for object colors. High chromas are preferred over low for object colors. The chief single factor responsible for pleasant color harmonies was found to be lightness contrast between object and background colors. The greater the lightness contrast, the greater are the chances of pleasant color combinations; this can be because lightness contrast is also most important for pattern vision. The influences of hue and saturation difference cannot be stated simply. A certain amount of variety, change, differentiation, or contrast is pleasant; sameness,

monotony, and repetition tend to be unpleasant. Configurations of colors should contain some variations in hue, lightness, and saturation, and over a period of time different configurations of colors should be employed to prevent satiation by overly familiar patterns of stimulation. Color preferences can differ due to such factors as function, size, configuration, climate, and sociocultural background.

Safety Colors Safety colors are used to indicate the presence of a hazard or safety facility such as an explosive hazard or a first aid station. These are carefully developed colors that are specified in American National Standard Z535.1-1998.51 The background around these safety colors should be kept as free of competing colors as possible, and the number of other colors in the area should be kept to a minimum. These colors should be illuminated by a light source to levels that both will permit positive identification of the color and the hazard or situation that it identifies and will not distort it and thereby obscure the message it conveys.

Figure 4-21. Specification of ANSI Safety Colors Viewed Under CIE Standard Illuminant C The specification of these colors is given in Figure 4-21. Designers must be aware that these color specifications are based on illuminant C. The colors will be recognizable under daylight and conventional incandescent and fluorescent light sources, which have a broad spectrum. High-intensity discharge light sources render some colors differently than the sources mentioned above. They can cause some confusion, especially at illuminances of 5 lx (0.5 fc) and lower, which are not uncommon in industrial spaces. Possible solutions are given in References 52 and 53. Color tolerance charts showing the safety colors and their tolerance limits are available from Hale Color Consultants.

Chromaticity and Illuminance The chromaticity of the light source should be matched to the illuminances.54 From experience it has been found that at low illuminances a "warm light" (less than 3300 K) is usually preferred, but the color temperature of the light source should increase as the illuminance increases.54 Recent data, however, both support and contradict this assertion, so these statements cannot be taken as conclusive.

COLOR RENDERING As previously discussed, lamps cannot be assessed for color rendering properties by visual inspection of the lamps themselves. To provide a color rendering index (CRI), it is necessary to have accurate and precise spectroradiometric measurements of light sources (see Chapter 2, Measurement of Light and other Radiant Energy). It is also necessary to understand the mechanisms of color vision, particularly chromatic adaptation. Knowledge in this area is still incomplete. However, in most cases it is possible to provide a useful answer. The recommendations below are based on the following assumptions. The color shift that occurs when an object is observed under different light sources can be classified in three ways: as a colorimetric shift, an adaptive shift, or a resultant color shift in which the first two are combined. To understand the subject it is extremely important that the three concepts be understood: 1. Colorimetric shift is the difference between the color (luminance and chromaticity) of an object illuminated by a nonstandard source and the color of the same object illuminated by the standard source, usually measured on a scale appropriate for assessing color differences. 2. Adaptive color shift is the difference in the perceived color of an object caused solely by chromatic adaptation. 3. Resultant color shift is the difference between the perceived color of an object illuminated by a nonstandard source and that of the same object illuminated by the standard source for specified viewing conditions. The conditions usually are that the observer shall have normal color vision and be adapted to the environment illuminated by each source in turn. The color shift is the resultant of the colorimetric and adaptive color shifts. The colorimetric shift can be determined using standard CIE conventions, but determination of the adaptive shift requires some assumptions about the effects of chromatic adaptation.

CIE Test-Color Method The CIE recommends a test-color method for measuring and specifying the color rendering properties of light sources.1 It rates lamps in terms of a color rendering index (CRI) that represents the degree of resultant color shift of a test object under a test lamp in comparison with its color under a standard lamp of the same correlated color temperature. The indices are based on a general comparison of the lengths of chromaticity-difference vectors in the 1964 uniform color space. The rating consists of a general index, Ra, which is the mean of the special indices, Ri, for a set of eight test-color samples that have been found adequate to cover the hue circuit (Plate 6). This can be supplemented by special indices based on such special test samples as CIE colors 9 to 14. Unless otherwise specified, the reference light source for sources with a correlated color temperature below 5000 K is a planckian radiator of the same correlated color temperature (Figure 4-14). For 5000 K and above, the reference source is one of a series of spectral energy distributions of daylight based on reconstituted daylight data30 developed from daylight measurements made in Enfield, England; Rochester, N.Y.; and Ottawa, Canada. Tables of colorimetric data are included in the CIE recommendations for planckian radiators up to 5000 K, and on these reconstituted daylight curves from 5000 K to infinity, for eight general and six special test-color samples. The current version of the CIE method is basically the same as an earlier version but with a better correction for the adaptive color shift. A paper by Nickerson and Jerome55 on the earlier version provides a working text and formulas, discusses the meaning of the index, and shows applications to a number of lamps. A 1962 IES report4 discusses in more detail the problems involved in the more-than-ten-year study of the subject. It indicates some of the problems, particularly those of chromatic adaptation, that remain to be solved before an all-purpose, completely satisfactory method can be established for rating a lamp, regardless of its color, against a single standard (probably daylight). Because of these problems, the index is not an absolute figure. For example, a 6500-K daylight lamp and a 3000-K warm white lamp having equal values on the general color rendering index will differ from their respective reference illuminants, the CIE phase of

daylight D65 and the 3000-K planckian radiator by approximately the same amount. These reference illuminants differ from one another in their color rendering, and so will the two lamps tested, even though they have the same general color rendering index.1 Figure 4-22 shows the basis for the CIE index where the test and reference lamp have the same chromaticity. CIE ratings are in terms of a single index, Ra, but to provide more information on the color rendering properties of a lamp, it is recommended that this be accompanied by a listing of the eight special index values on which the rating is based. Since the eight test samples cover the hue circuit, this makes it possible to obtain a record of the relative colorimetric shift in the different hues under the test lamp. Plotting the chromaticity difference vectors provides even more information, for this indicates the direction as well as the degree of colorimetric shift that is involved. If two lamps differ in Ri by approximately 5 units, the colors of test sample i rendered by the two lamps will be just perceptibly different under the best conditions, provided that the directions of the color shifts are nearly the same. No such simple rule can be given for Ra. It is obtained as the mean of eight Ri values, and even when two lamps have exactly the same Ra, differences of approximately 5 units or more in one or more of the Ri values can still be possible, so that their color rendering properties will be different for the object colors in question. Where the Ra values are close to 100, the Ri values are unlikely to show variations large enough to result in noticeable color differences. But as the value of Ra decreases from 100, the corresponding special indices Ri show increasing spread. Ratings are illustrated in Figure 4-23 for a number of typical lamps. The best color rendering lamps not only have a high index but also have the least variation in special indices for the different hues that are used as test samples. The closer one comes to a perfect color matching lamp, the tighter must be these tolerances.

Figure 4-22. Graphic basis for color rendering index. It makes sense to compare color rendering indices only of lamps with chromaticities close to each other. The standard illuminants A and D65 both have CRIs of 100 yet render colors quite differently. Likewise, commercial daylight and warm white lamps will render colors differently even if their CRI values match. The CIE Technical Committee TC 1-33 might change its recommendations for computing CRI in the future. The present method is based on the CIE 1964 uniform color space, which is now obsolete. In any case, it will eventually be more appropriate to change the basis of color rendering to a color appearance model.

Visual Clarity The concept of "visual clarity" has been used in several studies56,57 to indicate a preferred appearance of scenes containing colored objects when illuminated by certain sources. "Visual clarity" seems to be a combination of various factors including perceived color and contrast, color rendering, color discrimination, color preference, and border sharpness, but it is not as yet a well-understood notion.

LIGHT SOURCES FOR COLOR APPRAISAL, COLOR MATCHING, AND COLOR REPRODUCTION General Principles General lighting can be unsatisfactory for the precise appraisal of the colors of objects, including matching and reproducing colors. Such tasks are necessary in industries that make and market pigments and dyes, and in arts and industries involving color production (e.g., painting, textile dyeing, photography, and color printing). These tasks are also necessary in commercial evaluation of such naturally colored objects as fibers, foods, minerals, and gemstones. Although color processes often are controlled by means of instrumentation, the ultimate approval of colored products is based on visual judgment, generally the comparison of the product with a colored standard. The grade of a natural material can be based on a visual judgment of the correspondence of its colors to a series of standards.

Figure 4-23. Color and Color Rendering Characteristics of Common Light Sources Since two objects might match under one kind of illumination but not match under another, the kind of illumination used for inspection is crucial. The specification of the illuminant must be expressed in any written or oral contract or other commitment involving color production. Very often the intention is to have the product match the standard satisfactorily under any kind of illumination likely to be encountered. This requires that the product be compared with the standard under several kinds of illumination, typically two phases of daylight of widely different correlated color temperature, as well as fluorescent and incandescent light. Natural daylight is highly variable, not available at night, and not easily available in interior spaces, so a source of electric light that simulates daylight is usually used. Clearly, lighting for these purposes must meet specifications far more stringent than those that are applied to general interior and exterior lighting, and these specifications have been agreed to by national and international standardizing bodies. Fortunately for the practicing illuminating engineer, the specifications are easily met because viewing booths that exclude extraneous light and provide several kinds of illumination meeting these standards are commercially available and are customarily used in these applications. Light used for these purposes must have a specified SPD. Because of the widespread natural occurrence and deliberate use of fluorescence, the ultraviolet spectrum must be specified as well as the visible spectrum. Other, less restrictive means of specifying light are not satisfactory. A given correlated color temperature permits an infinite variety of chromaticities and a given chromaticity permits an infinite variety of spectra. The spectral quality of light sources that simulate standard daylight for judging colors is assessed by a method adopted by the CIE.58 The basic premise of this method is that the best of several available sources is the one that causes the least average color difference for specified pairs of hypothetical colored surfaces that match metamerically in daylight. In some cases, especially in the grading of natural materials, the color difference between grades can be accentuated and made more readily apparent by a wellchosen light source. For example, if materials yellow with age, yellowness can be regarded as undesirable. Slight differences in yellowness can be more readily perceived with a light source that appears white but is rich in the short-wavelength end of the spectrum. Illuminance affects color judgment and must be specified. Dark colors require more illuminance than light colors. Some materials, notably wood and textiles as well as metallic and pearlescent paints, might match under certain angular conditions of illumination and viewing, but not under others. For this reason, the geometry of illumination and viewing must be specified. In most specifications, opaque surfaces are illuminated at 45° to their normal and viewed on the normal, or the reverse of this arrangement. The background or surround color influences color judgments, so it must be specified. Small color differences are best perceived if the surround is of a color between the two colors being compared. This ideal is usually approximated by the use of a neutral (black, gray or white) surround of approximately the same lightness as the specimens. A neutral surround does not influence the appearance of hue. When glossy specimens are viewed at 45° to their normal, something on the opposite side of the normal can be seen reflected in the surfaces. That specularly reflected light is added to the diffusely reflected light, interfering with color judgment. It is standard practice to minimize this effect by placing black velvet on the opposite side. Color differences usually are controlled through the use of a set of seven colors that constitutes a color tolerance standard. The set utilizes an ideal color, the lightest and darkest acceptable colors of the same hue and chroma, the two extremes of acceptable hue variation with the same lightness and chroma, and the weak and strong (chroma) limits with the same lightness and hue. The colors are arranged on a card with slots permitting the direct comparison of the underlying surface with each of the standard colors.

Standard Viewing Conditions The grading of raw cotton, on the basis of its color, is a visual task of great commercial importance. Studies in the U.S. Department of Agriculture, reported as early as 1939, provided the basis for specifying 600 to 800 lx (60 to 80 fc) and a SPD closely simulating daylight, with a correlated color temperature of 7500

K, for this task. To minimize glare, the geometric relationship of the luminaires to the work surface is also specified.59 Recommendations for viewing textiles have been made by the American Association of Textile Chemists and Colorists, and for judging diamonds by the Gemological Institute of America. The results of experiments coordinated by a committee of the Inter-Society Color Council indicated that textile color matchers prefer a range of correlated color temperatures that depends on the illuminance, as shown in Figure 4-24.60 Conditions for comparison of opaque specimens, in general, have been standardized by the American Society for Testing and Materials (ASTM).61 Several different SPD are specified, so proposed color matches can be examined under various lights to test for or optimize metamerism. In addition to daylight, a light simulating sunlight, a typical tungsten source, and cool white fluorescent lamps are specified. For materials of medium lightness, an illuminance of 1000 to 1250 lx (100 to 125 fc) is recommended for critical evaluation, and 750 to 1750 lx (75 to 175 fc) for general evaluation. For very light materials, 500 lx (50 fc) is adequate, and for very dark materials, the illuminance can be increased to as much as 2000 lx (200 fc). Specifications for viewing conditions in photography and graphic arts have been standardized by the American National Standards Institute (ANSI). Efforts in improving these specifications are ongoing.62 In these applications, it is standard practice to use illumination simulating a phase of daylight having a correlated color temperature of 5000 K. Large transparencies are viewed against a luminous surface of standard luminance and spectral distribution, and small transparencies are viewed by means of a standardized projection system. Both methods permit direct comparison of transparencies and opaque prints.

Figure 4-24. Tests conducted under the direction of the Inter-Society Color Council show the characteristics of preferred daylight illumination conditions for color matching, grading, and classing. In the United States, standards other than those for photography and color printing specify a phase of daylight having a correlated color temperature of 7500 K, typical of light from a slightly overcast north sky (south sky in the southern hemisphere). Most of the rest of the world follows a recommendation of the CIE to use 6500 K, which corresponds to average daylight. Viewing booths of both kinds are commercially available.

Color in Imaging Technology Color photography is based on the same general principle as color vision: the eye employs three kinds of cones, and color film has three photosensitive layers, each maximally sensitive to a different part of the visible spectrum. Films designed for making transparencies are different from those designed for making opaque prints, but the same three-layer principle is used. The spectral sensitivities of the three film layers differ considerably from those of the human visual system, and the dye images are at best metameric matches to the objects photographed. Objects that match when viewed directly might not match in the photograph, and objects that do not match when viewed directly might match in the photograph. Color films are designed to give satisfactory color rendition when the subject is illuminated with a specific kind of source. Flash lamps, for example, are designed to simulate daylight. ANSI and the International Organization for Standardization (ISO) have adopted standards specifying light sources for testing films, the color contribution of camera lenses, and methods for testing flash equipment, as well as conditions for viewing prints and transparencies.63-67 Television images are mediated through the spectral sensitivities of the camera, the electronic processing of the image, and the three color phosphors on the face of the display tube. The problems associated with color reproduction in photography are also experienced in television, with the added burden of a far greater variety in the characteristics of the final display medium. The quality of the color reproduction depends, of course, on the spectral quality of the illumination of the scene to be televised, as well as the characteristics of the equipment. In color printing, various kinds of plates are used to apply colored inks to paper. The inks are usually yellow, magenta, cyan, and black. The plates are made by photoengraving or by electronic scanners. Like color photography, color printing also can suffer from discrepancies in color appearance with the actual objects. The color rendition of color reproduction processes can be evaluated by the use of a commercially available colored test chart designed for this purpose.68 The chart is photographed, viewed by television equipment, or printed, and the resulting image is compared with the original chart or a second chart of the same kind, under standardized viewing conditions. The comparison can be made quantitative by performing color difference measurements. These charts also can be used to make simple direct visual appraisals of the color rendering of light sources. The CIE recognizes that lighting is critical for the appearance of colors displayed by CRTs and printers. The CIE has established a new division (Division 8 Image Technology) to address these important issues.

REFERENCES 1. Commission Internationale de l'Éclairage. CIE Committee TC-3.2. 1995. Method of measuring and specifying colour rendering properties of light sources, CIE no. 13.3. Paris: Bureau Central de la CIE. 2. IES. Testing Procedures Committee. Photometry of Light Sources Subcommittee. 1989. IES practical guide to colorimetry of light sources, IES LM-16-1984. J. Illum. Eng. Soc. 18(2):122-127. 3. IES. Color Committee. 1962. Color and the use of color by the illuminating engineer. Illum. Eng. 57(12):764-776. 4. IES. Light Sources Committee. Subcommittee on Color Rendering. 1962. Interim method of measuring and specifying color rendering of light sources. Illum. Eng. 57(7):471-495. 5. IES. Color Committee. 1992. Color and illumination, IES DG-1-1990. New York: Illuminating Engineering Society of North America. 6. Judd, D. B., and G. Wyszecki. 1975. Color in business, science and industry. 3rd ed. New York: John Wiley. 7. Nickerson, D. 1960. Light sources and color rendering. J. Opt. Soc. Am. 50(1):57-69. 8. Hunt, R. W. G. 1978. Colour terminology. Color Res. Appl. 3(2):79-87. 9. Newhall, S. M., D. Nickerson, and D. B. Judd. 1943. Final report of the OSA subcommittee on the spacing of the Munsell colors. J. Opt. Soc. Am. 33(7):385418. 10. Commission Internationale de l'Éclairage. 1986. Colorimetry, CIE no. 15.2. Paris: Bureau Central de la CIE. 11. Robertson, A. R. 1977. The CIE 1976 color-difference formulae. Color Res. Appl. 2(1):7-11. 12. American Society for Testing and Materials. 1997. Method E308, computing colors of objects by using the CIE scale. ASTM E308-96. West

Conshohocken, PA: ASTM. 13. Wyszecki, G., and W. S. Stiles. 1982. Color science: Concepts and methods, quantitative data and formulae. 2nd ed. New York: John Wiley. 14. MacAdam, D. L. 1937. Projective transformations of ICI color specifications. J. Opt. Soc. Am. 27(8):294-299. 15. Wyszecki, G. 1963. Proposal for a new color-difference formula. J. Opt. Soc. Am. 53(11):1318-1319. 16. Recommendations on uniform color spaces, color-difference equations, and metric color terms. 1977. Color Res. Appl. 2(1): 5-6. 17. Clarke, F. J. J., R. McDonald, and B. Rigg. 1984. Modification to the JPC79 colour difference formula. J. Soc. Dyers Col. 100(4):128-132. 18. American Association of Textile Chemists and Colorists. 1992. CMC: Calculation of small color differences for acceptability, Test Method 173-1992. Research Triangle Park, NC: American Association of Textile Chemists and Colorists. 19. American Society for Testing and Materials. 1996. Standard practice for specifying color by the Munsell system, D1535-96. Philadelphia: ASTM. 20. Kelly, K. L., and D. B. Judd. 1965. The ISCC-NBS centroid color charts, SRM #2106. Washington: Office of Standard Reference Materials, National Bureau of Standards. 21. Kelly, K. L. 1965. The universal color language. Color Eng. 3(2):16-21. 22. National Bureau of Standards. 1976. Color: Universal language and dictionary of names, prepared by Kenneth L. Kelly, and Deane B. Judd. NBS Special Publication 440. Washington: U. S. Government Printing Office. 23. MacAdam, D. L. 1974. Uniform color scales. J. Opt. Soc. Am. 64(12):1691-1702. 24. Hård, A., and L. Sivik. 1981. NCS--Natural Color System: A Swedish standard for color notation. Color Res. Appl. 6(3): 127-138. 25. Richter, M., and K. Witt. 1986. The story of the DIN color system. Color Res. Appl. 11(2):138-145. 26. Nickerson, D. 1957. Horticultural color chart names with Munsell key. J. Opt. Soc. Am. 47(7):619-621. 27. Billmeyer Jr., F. W. 1987. Survey of color order systems. Color Res. Appl. 12(4):173-186. 28. Kelly, K. L. 1963. Lines of constant correlated color temperature based on MacAdam's (u,v) uniform chromaticity transformation of the CIE diagram. J. Opt. Soc. Am. 53(8):999-1002. 29. Robertson, A. R. 1968. Computation of correlated color temperature and distribution temperature. J. Opt. Soc. Am. 58(11): 1528-1535. 30. Judd, D. B., D. L. MacAdam, and G. Wyszecki. 1964. Spectral distribution of typical daylight as a function of correlated color temperature. J. Opt. Soc. Am. 54(8):1031-1040. 31. Evans, R. M. 1948. An introduction to color. New York: John Wiley. 32. Burnham, R. W., R. M. Hanes, and C. J. Bartleson. 1963. Color: A guide to basic facts and concepts. New York: John Wiley. 33. Hunt, R. W. G. 1991. A revised colour appearance model for related and unrelated colours. Color Res. Appl. 16(3):146-165. 34. Nayatani, Y., K. Takhama, H. Sobagaki, and K. Hashimoto. 1990. Color-appearance model and chromatic adaptation transform. Color Res. Appl. 15(4):210221. 35. Commission Internationale de l'Éclairage. 1998. The CIE 1997 interim colour appearance model (simple version), CIE no. 131. Vienna: Bureau Central de la CIE. 36. Stiles, W. S., and G. Wyszecki. 1968. Intersections of the spectral reflectance curves of metameric object colors. J. Opt. Soc. Am. 58(1):32-40. 37. Brainerd, A. A., and R. A. Massey. 1942. Salvaging waste light for victory. Illum. Eng. 37(10):738-757. 38. Moon, P. 1941. Wall materials and lighting. J. Opt. Soc. Am. 31(12):723-729. 39. Spencer, D. E., and S. E. Sanborn. 1961. Interflections and color. J. Franklin Inst. 252(5):413-426. 40. O'Brien, P. F. 1960. Lighting calculations for thirty-five thousand rooms. Illum. Eng. 55(4):215-226. 41. Jones, B. F., and J. R. Jones. 1959. A versatile method of calculating illumination and brightness. Illum. Eng. 54(2): 113-121. 42. Barthès, E. 1957. Études expérimentales de calcul de point de couler de la lumière reçue par le plan utile dans un local à parois colorées. Bul. Soc. Fran. Elec. 7(81):546-542. 43. Krossawa, R. 1963. Color shift of room interior surfaces due to interreflection. Die Farbe 12:117. 44. Yamanaka, T., and Y. Nayatani. 1964. A note of predetermination of color shift due to interreflection in a colored room. Acta Chromatica 1:111. 45. Helson, H. 1954. Color and vision. Illum. Eng. 49(2):92-93. 46. Helson, H. 1955. Color and seeing. Illum. Eng. 50(6): 271-278. 47. Helson, H., D. B. Judd, and M. Wilson. 1956. Color rendition with fluorescent sources of illumination. Illum. Eng. 51(4): 329-346. 48. Helson, H. 1965. Role of sources and backgrounds on pleasantness of object colors. IES National Technical Conference, New York. 49. Judd, D. B. 1971. Choosing pleasant color combinations. 1(2): 31-41.

50. Helson, H., and T. Lansford. 1970. The role of spectral energy of source and background color in the pleasantness of object colors. Appl. Opt. 9(7):15131562. 51. American National Standards Institute. 1998. Safety color code, ANSI Z535.1-1998. New York: ANSI. Illuminating Engineering Society. Color Committee. 1980. Potential misidentification of industrial safety colors with certain lighting. Light. Des. Appl. 10(11):20. 52. National Bureau of Standards. 1983. Some criteria for colors and signs in workplaces, prepared by Robert A. Glass, Gerald L. Howett, Karen Lister, and Belinda L. Collins. NBSIR 83-2694. Washington: National Bureau of Standards. 53. Commission Internationale de l'Éclairage. 1986. Guide on interior lighting, CIE no. 29.2. Paris: Bureau Central de la CIE. 54. Kruithof, A. A. 1941. Tubular luminescence lamps for general illumination. Philips Tech. Rev. 6(3):65-73. 55. Nickerson, D., and C. W. Jerome. 1965. Color rendering of light sources: CIE method of specification and its application. Illum. Eng. 60(4):262-271. 56. Aston, S. M., and H. E. Bellchambers. 1969. Illumination, colour rendering and visual clarity. Light. Res. Tech. 1(4):259-261. 57. Thornton, W. A., and E. Chen. 1978. What is visual clarity? J. Illum. Eng. Soc. 7(2):85-94. 58. Commission Internationale de l'Éclairage. 1981. A method for assessing the quality of daylight simulators for colorimetry. CIE no. 51. Vienna: Bureau Central de la CIE. 59. American Society for Testing and Materials. 1996. Standard practice for lighting cotton classing rooms for color grading, ASTM D1684-96. Philadelphia: ASTM. 60. Nickerson, D. 1948. The illuminant in textile color matching. Illum. Eng. 43(4):416-467. 61. American Society for Testing and Materials. 1996. Standard practice for visual appriasal of colors and color differences of diffusely-illuminated opaque materials, D1729-96. Philadelphia: ASTM. 62. Johnson, T. 1996. Colour appearance - Standardization of viewing conditions and measurement procedures. In Proceedings of the CIE Symposium '96 on Color Standards for Image Technology. Vienna: Bureau Central de la CIE. 63. Amphoto. 1978. Lighting. Vol. 9 in Encyclopedia of practical photography. Garden City, NY: American Photographic Book Publishing Co. 64. Society of Photographic Scientists and Engineers. 1973. SPSE handbook for photographic science and engineering, edited by Thomas Woodlief. New York: John Wiley. 65. Spencer, D. A. 1966. Colour photography in practice. Rev. ed. London, New York: Focal Press. 66. Hunt, R. W. G. 1967. The reproduction of colour. 2nd ed. London: Fountain Press. 67. McCamy, C. S. 1959. A nomograph for selecting light balancing filters for camera exposure of color films. Photogr. Sci. Eng. 3(6):302-304. 68. McCamy, C. S., H. Marcus, and J. G. Davidson. 1976. A color-rendition chart. J. Appl. Photo. Eng. 2(3):95-99.

5 Nonvisual Effects of Optical Radiation Humans, animals, and plants have complex physiological responses to the daily and seasonal variations in solar radiation under which they evolved. The spectral power distribution of solar radiation is shown in Figure 5-1. The study of the interaction of biological systems with nonionizing radiant energy in the ultraviolet (UV), visible, and infrared (IR) portions of the electromagnetic spectrum is known as photobiology. Photobiological responses result from chemical and physical changes induced by the quantal absorption of radiation by specific molecules in the living organism. The absorbed radiation produces excited electronic states in these molecules, which can lead to photochemical reactions of biological consequence. The distinguishing feature of photochemical reactions is that the activation energy is provided by the quantum absorption of nonionizing photons, which cause reactions to occur at low (physiological) temperatures. The human visual system is discussed in Chapter 3, Vision and Perception. Extravisual photobiological effects occurring in humans, animals, microorganisms, and plants, as well as nonbiological effects on matter, are covered in this chapter. The term "optical radiation" refers to the wavelength of the radiant energy described in this chapter (between 100 nm and 1 mm), primarily consisting of UV, visible, and IR radiation.

EFFECTS ON HUMANS AND ANIMALS (PHOTOBIOLOGY) The effects of solar radiation on humans and animals include such wide-ranging phenomena as damage to ocular tissues, skin effects, tumor formation, and the synchronization of biological rhythms. A variety of diseases have been treated with visible or UV energy, alone or in combination with sensitizing drugs. Since the beginning of recorded history, psoralen, a UV-activated drug, combined with exposure to solar UV radiation, has been the therapy for vitiligo, a skin condition marked by an absence of normal pigment. By the turn of the century, lupus vulgaris, a condition where skin nodules are present, was shown to be cured with UV either from sunlight or from carbon arcs. Psoriasis, a skin condition where lesions are covered with scales, is now being alleviated with the same therapy that has been applied to vitiligo, using electric light sources of more constant UV output than from the sun. Visible radiation, particularly the short wavelengths (400 to 500 nm), is used in the phototherapy of jaundiced infants. Photodynamic therapy (PDT) is a rapidly developing field for the treatment of such disorders as cancers involving rapid cell division. A photoactive dye is absorbed by dividing cells. When the dye is photoactivated in the presence of oxygen, free radicals develop that critically damage the reproducing cells. Other extravisual effects of light include the regulation of biological rhythm and neuroendocrine responses.

Effects on the Eye1-38 For the purposes of this discussion, the optical radiation spectrum is divided into three components: UV, 100 to 380 nm; visible and near-IR, 380 to 1400 nm; and IR, 1400 nm to 1 mm. Figure 5-2 summarizes in abbreviated form the overall effects of radiation as a function of wavelength and indicates that UV bands, in particular, induce such adverse effects as actinic erythema (reddening of the skin), photokeratitis (an inflammation of the cornea, also commonly known as "flash blindness" or "welder's burn"), and photosensitized skin damage, as well as some beneficial effects, as in phototherapy. Three elements are involved in optical radiation damage to various components of the eye: the accessibility of a given wavelength to the tissue in question, the absorbance of that wavelength, and the ability of the tissue to deal with the insult that the absorption of energy represents. UV Radiation Effects. The UV region of the electromagnetic spectrum is subdivided by the Commission International de l'Éclairage (CIE) into near (UV-A, 315 to 400 nm), middle (UV-B, 280 to 315 nm) and far (UV-C, 100 to 280 nm) UV bands. Figure 5-3 shows that for wavelengths less than 320 nm, nearly all of the radiation is absorbed by the cornea. Between 320 and 400 nm, much of the UV radiation is absorbed by the lens; the proportion is dependent on age (Figure 5-4). The optical media of the human eye, until early adulthood, transmit a small percentage of UV radiation to the retina, resulting in a theoretical visual response for wavelengths as short as 300 nm.

Figure 5-1a. Spectral power distribution of solar radiation at sea level, showing the ozone, oxygen, water, and carbon dioxide absorption bands.

Figure 5-1b. Global irradiance, solar noon at Durham, UK (55° N) : (a) 2 October 1986 (uniform light cloud); (b) 1 July 1986 (clear sky). Photokeratitis is a painful but not necessarily serious inflammation of the epithelial (outermost) layer of the cornea. The period of latency between exposure and the onset of symptoms varies from 2 to 8 hours, depending on the amount of radiation received. For moderate exposures, the effects are more frightening than serious. The symptoms include inflammation of the conjunctiva accompanied by a reddening of the surrounding skin and eyelids. There is a sensation of sand in the eyes, tearing, sensitivity to light, and twitching of the eyelids. Recovery is rapid and usually complete within 48 hours except for severe cases. The action spectrum, similar to that for skin erythema, peaks at 270 to 280 nm (with recent research suggesting that it is closer to 270 nm) (Figure 5-5). Lenticular effects from UV radiation recently have been undergoing extensive investigation. The lens shows a number of changes with aging, including a yellowing coloration, an increasing proportion of insoluble proteins, sclerosis with loss of accommodation, and cataract. There is a growing body of evidence, mostly epidemiological, to implicate UV radiation in these changes. For example, cataract extractions are significantly more frequent in India than in western Europe. Part of the difference may be due to diet and genetic factors, but most authorities believe that exposure to sunlight plays an important role. While many of the early epidemiological studies of cataract have been inconclusive, more recent attempts have shown statistical significance in the relationship between cortical lens opacities and lifelong UV-B exposure in persons living and working in high levels of solar energy.35 Suggestions have been made that UV-A also may have a role in cataract formation. There are arguments21 that UV exposure might not be a significant causal factor for cataracts. Until these issues are resolved, the conservative approach is to minimize unnecessary UV exposure of the eyes. Retinal effects of UV radiation are difficult to categorize because they depend on the individual filtering capabilities of the preretinal ocular media. In adults, the crystalline lens, which typically absorbs wavelengths below about 400 nm, effectively shields the retina from UV radiation. Studies have shown, however, that a small percentage of UV radiation can reach the retina in human adults up to 30 years of age. Removal of the lens in cataract surgery renders the retina more susceptible to damage from wavelengths down to 300 nm. If a UV-blocking intraocular lens (IOL) is surgically implanted, however, then the UV absorption is restored. UV shielding is also available for rigid gas-permeable (RGP) and hydrogel varieties of contact lenses. Data on UV radiation from some common sources can be found in the "Museum" section of Chapter 14, Lighting for Public Places and Institutions.

Figure 5-2. Physiological Effects or Applications of Ultraviolet, Visible and Near-Infrared, and Infrared Radiation

Figure 5-3. Percentage of energy on the surface of the cornea absorbed by various layers in the normal adult eye. Visible and Near-IR Effects. The IR region of the electromagnetic spectrum has been divided into three subregions: IRA (near-IR, 780 to 1400 nm), IR-B (middle-IR, 1400 to 3000 nm), and IR-C (far-IR, 3000 nm to 1 mm) bands. Visible radiation occupies the wavelength region bounded by UV and IR, falling between approximately 400 and 780 nm. While the sun radiates at all wavelengths across the electromagnetic spectrum, only UV, IR, and visible radiation reach the earth's surface. Retinal injury resulting in a loss of vision (scotoma) following observation of the sun has been described throughout history. The incidence of chorioretinal injuries from fabricated light sources is extremely small and is no doubt far less than the incidence of eclipse blindness. Until recently, chorioretinal burns resulting from industrial

operations were rare occurrences. Indeed, this is still largely true, since the normal aversion to high-brightness light sources (the blink reflex and movement of the eyes away from the source) provides adequate protection unless the exposure is hazardous within the duration of the blink reflex. The recent revolution in optical technology, however, forged principally by the invention of the laser, has meant a great increase in the use of high-intensity, high-radiance sources. Many such sources have output parameters significantly different from those encountered in the past and may present serious chorioretinal burn hazards. In addition to lasers, one may encounter the following sources of continuous optical radiation in industry: compact arc lamps (as in solar simulators), tungsten-halogen lamps, gas and vapor discharge tubes, electric welding units, and sources of pulsed optical radiation, such as flash lamps and exploding wires. The intensities of these sources may be of concern if adequate protective measures are not taken. Extreme IR irradiances have been linked to corneal, lenticular, and retinal damage; although the ocular structures can adequately dissipate the heat from low-power diffuse IR exposures, the same amount of energy delivered in pulses to very small areas of tissue can cause damage. Coherent light generated by yttrium aluminum garnet (YAG) and argon lasers can penetrate to intraocular structures. Light from krypton, HeNe, and ruby lasers can reach the retina. Such sources have been used therapeutically in retinal photocoagulation procedures.

Figure 5-4. Human lens average transmittance.

Figure 5-5. Comparison of the radiant exposure thresholds for injuring the cornea of the human, the primate and the rabbit. The data were established by exposing 238 rabbit eyes, 83 primate eyes, and 39 human eyes to ultraviolet radiation. To place chorioretinal injury data in perspective, Figure 5-6 shows the retinal irradiance for many light sources. It is reemphasized that several orders of magnitude in radiance or luminance exist between sources that cause chorioretinal burns and those levels to which individuals are continuously exposed. The retinal irradiances shown are only approximate and assume minimal pupil sizes and some squinting for the very high luminance sources (except the xenon searchlight, for which a 7-mm-diameter pupil was assumed so as to apply to nighttime illumination). Light entering the cornea passes through the anterior chamber (which contains the aqueous humor), the lens, and the vitreous humor and impinges upon the retina (Figure 3-1 in Chapter 3, Vision and Perception). Examination of Figure 3-6 shows that between 400 and 1400 nm the retina is vulnerable to radiation effects. Between these wavelengths the retina is by far the most sensitive tissue of the body.

Figure 5-6. The eye is exposed to light sources having radiances varying from about 104 to 106 W/(cm-2 · sr) or less. The resulting retinal irradiances vary from about 200 down to 10-7 W/cm-2 and even lower; retinal irradiances are shown for typical image sizes for several sources. A minimal pupil size was assumed for intense sources, except for the searchlight. The retinal burn threshold for a 10-s exposure of the rabbit retina is shown as the upper solid line. The maximum permissible exposure (MPE) applied by the U.S. Army Environmental Hygiene Agency in evaluating momentary viewing of continuous-wave light sources is shown as the lower solid line. Approximate pupil sizes are shown at the lower left, based upon exposure of most of the retina to light of the given irradiance. In the retina, light passes through multiple layers of neural cells before encountering the photoreceptor cells (the rods and cones). Photoreceptors are neural transducers, converting absorbed photons of light into electrical impulses sent to the brain via the optic tract. Just behind the rods and cones is a single layer of heavily pigmented cells (the pigment epithelium), which absorbs a large portion of the light passing through the neural retina. The pigment epithelium acts like a dark curtain to absorb and prevent backscatter from those photons that are not absorbed in the outer segments of the rods and cones. The neural retina itself is almost transparent to light. The pigment epithelium is approximately 10 µm thick, while the choroid, a layer of blood vessels behind it, ranges in thickness from 100 to 200 µm. Most of the light that reaches the retina is converted to heat by the pigment epithelium and the choroid. Sufficiently large quantities of light can generate sufficient heat to damage the retina. Research in recent decades has demonstrated that for radiant energy between 400 and 1400 nm, there are at least three different mechanisms leading to retinal damage. These are: 1. Thermal damage from pulse durations extending from microseconds to seconds. Except for minor variations in transmittance through the ocular media and variations of absorbance in the pigmented epithelium and choroid, thermal damage is not wavelength dependent. 2. Photochemical damage from exposure to short wavelengths in the visible spectrum for time durations and power densities on the retina that preclude thermal effects. Photochemical damage is wavelength dependent. 3. Mechanical (shock-wave) damage from picosecond and nanosecond pulses of mode-locked or Q-switched lasers.

In terms of exposure time and wavelength there is no abrupt transition from one type of damage to the other. For example, a YAG laser emitting a pulse train of Q-switched pulses (several ns in duration) at 1064 nm can produce a combination of shock-wave and thermal damage depending on the pulse width and the time interval between pulses, whereas an acoustically modulated pulse train from an argon ion laser emitting 10-µs pulses of 488-nm radiation might produce a combination of thermal and photochemical damage. A number of researchers have shown that long-term exposure to light can cause retinal damage in some animals. For example, when rats and mice are subjected to cool white fluorescent lighting for extended periods of time (weeks to months), they become blind. Histological examination reveals that the photoreceptors in the retinae of these animals have degenerated. Although rodent retinal photoreceptors can be damaged with long exposures to relatively low levels of white light,37,38 such damage in primates has been demonstrated only with the eyes dilated and at a continuous exposure of 10,800 lux for 12 hours. Exposure of the undilated monkey eye at that illuminance for 12 hours per day for 4 weeks did not produce photoreceptor damage.34 The role of light in the incidence of retinopathy of prematurity (ROP) in low-birthweight infants has been explored by several researchers,9,14,16 because these infants often are exposed to high levels of ambient or phototherapy light. One potential mechanism involves the ability of light of sufficient intensity to catalyze oxygen-producing chemical reactions at the retina.30,32 Elevated oxygen levels had been implicated as a risk factor for ROP in the 1950s.2 High intensities of light also increase the rate of blood flow in the ophthalmic artery,3 which also might elevate oxygen concentrations at the retina. A relationship between light and ROP remains to be demonstrated. Several studies failed to find a link between the two.1,29 Far-IR Effects. Very little IR radiation of wavelengths longer than 1400 nm reaches the retina (see Chapter 3, Vision and Perception), but such radiation can produce ocular effects leading to corneal and lenticular damage. Cataracts from exposure to IR radiation have been reported in the literature for a long time, but there are few and no recent data to substantiate the clinical observations. It was previously thought that long-term exposure to IR radiation produces an elevated temperature in the lens which, over a period of years, leads to denaturation of the lens proteins with consequent opacification. It is now believed that IR radiation is absorbed by the pigmented iris and converted to heat that is conducted to the lens, rather than by direct absorption of radiation in the lens. IR cataractogenesis has been reported to occur among glassblowers, steel puddlers, and others who undergo long-term occupational exposure to IR radiation. Present industrial safety practices have virtually eliminated this effect. Photosensitization. Retinal and other ocular effects also can be increased or decreased in severity by the presence of endogenously or exogenously supplied photoactive compounds. Psoralens, hematoporphyrin derivatives, and other phototherapeutic agents can enhance the damaging effects of various wavelengths on the eye and other tissues. In contrast, vitamin E can act as a quencher of excited-state and related species and has been hypothesized to increase the threshold for light-induced damage. Many new pharmaceutical agents contain conjugated bond and ring structures that also can increase the potential for phototoxic effects.

Effects of UV Radiation on the Skin39-56 There are at least two known benefits of UV radiation exposure on skin: the production of vitamin D from precursor chemicals, which are formed in the skin (see below), and the induction of protective pigmentation. Known harmful effects include sunburn, skin cancer, and morphologic alterations (wrinkling, irregularity, altered pigmentation, thinning and thickening of skin), which appear as premature aging. Delayed tanning and increased thickening of skin is a protective response initiated by UV radiation. The function of immediate tanning is uncertain. Optical Properties of the Skin. The reflectance of skin for wavelengths shorter than 320 nm is low, regardless of skin color; however, from 320 to 750 nm the reflectance is dependent on skin pigmentation. The transmission of UV radiation through the skin depends on wavelength, skin color (melanin content), and skin thickness. In general, transmission increases with increasing wavelength from 280 to 1200 nm. Typically, for those of European descent, the transmittance through the top layer of skin (stratum corneum) is 35% at 300 nm and 60% at 400 nm. In persons of African descent, the transmittance of the stratum corneum is about 20% at 300 nm and 40% at 400 nm. Transmission decreases with increasing melanin content of the skin and with increasing skin thickness (Figures 5-7 and 5-8).

Figure 5-7. Spectral transmission of two individuals, one with heavily pigmented skin and one with lightly pigmented skin. The dashed lines indicate the spectral transmission through the entire epidermis; the solid lines are for the top layer of the epidermis, the stratum corneum, alone.

Figure 5-8. Spectral transmission through the epidermis of equal thickness (12 µm). The dashed line is for one person with lightly pigmented skin; the solid line is for one person with heavily pigmented skin. While skin color is the genetically determined result of a number of factors, the primary factor is melanin. Melanin protects against UV damage by reducing transmission through absorption and scattering. Its quantity, granule size, and distribution all affect skin color. The immediate tanning that occurs with exposure to UV-A radiation and extending into the visible region is the darkening of existing melanin. Delayed tanning results from UV stimulation of the melaninproducing cells (the melanocytes) to produce additional melanin. Pigmentation from this process begins immediately at the subcellular level. Fading requires months, as melanin is lost during the normal shedding process. Erythema. The delayed reddening (actinic erythema) of the skin caused by exposure to UV radiation is a widely observed phenomenon. The spectral efficiency of this process (Figure 5-9a and b), particularly for sunlight radiation between 290 and 320 nm, has been well studied. The reported erythema action spectrum for wavelengths shorter than 290 nm varies considerably among observers because of differences in the degree of erythema taken as the endpoint criterion and differences in the time of observation after irradiation. In the past, no single erythemal action spectrum had been universally adopted. One based on the work of Coblentz42 has been used frequently, but subsequent studies have shown significant differences. In 1987, a reference erythemal spectrum was proposed by the CIE (Figure 5-9a and b),43 and it should supplant the various functions used in the past. Erythema is a component of skin inflammation and results from increased blood volume in superficial cutaneous vessels. Affected skin can therefore be warm and tender.

Figure 5-9a. Erythema action spectrum for human skin. Note logarithmic vertical axis.

Figure 5-9b. Values Used to Approximate Approximately 25 mJ/cm2 of energy at the most effective wavelength (297 nm) causes a barely perceptible reddening in fair-skinned caucasions.44 This amount of effective energy can be experienced during a 12-min exposure under overhead sun in the tropics where the stratospheric ozone layer is thinner. When the sun is 20° from its zenith and the ozone layer thickness is greater, an exposure of 20 min is typically required for the same degree of reddening. Exposure to UV radiation (particularly at high irradiance levels) can cause immediate erythema. Fading can occur a few minutes after irradiation ceases, and can reappear after 1 to 3 hours. The greater the dose, the faster the reappearance, and the longer the persistence of erythema. If the erythema is severe, skin peeling (desquamation) can begin approximately 10 days after exposure. This rapid sloughing off of the top skin layer results from the increased proliferation of skin cells during recovery after UV damage. Desquamation carries away some of the melanin granules stimulated by the UV radiation.

Photoprotection, in its common usage, refers to the protection against the detrimental effects of optical radiation afforded by sunscreens topically applied to the skin. These sunscreens reduce the effect of UV exposure primarily by absorption, but also by reflection in some cases. Some sunscreens are effective and relatively resistant to being washed away by sweating or swimming. Paraaminobenzoic acid (PABA) in an alcohol base has proven quite effective in preventing sunburn. Other materials in use include benzophenones, cinnamates, and salicylates. Effects of Dermal Radiation on the Immune System. Photoimmunology is the study of nonionizing radiation, predominantly in the UV portion of the spectrum, on the immune system. The photoimmunologic effects of UV radiation are selective: only a few immune responses are affected. The alterations studied in greatest detail are the induction of susceptibility to UV-induced neoplasia and systemic and local suppression of contact hypersensitivity. Most observations have been made in experimental animal systems, although some photoimmunologic effects have been observed in humans. UV radiation can affect immunity systematically. For example, exposure of the skin to UV at one place on the body can reduce the sensitivity to UV at unexposed sites. This probably occurs through the release of mediators from the skin at the exposure site, which in turn results in the formation of antigen-specific T suppressor lymphocytes (white blood cells); such cells have been found in the spleens of animals.

Figure 5-10. Steps in the photoproduction of vitamin D in the skin. Skin Cancer. The three varieties of skin cancer are basal cell, squamous cell, and malignant melanoma. The frequency of occurrence is in the order stated, basal cell cancer being the most common. The prevalence of basal cell carcinoma varies inversely with latitude. The prevalence of both basal and squamous cell cancer correlates positively with solar UV exposure, but there is some evidence that UV exposure after age 10 might not contribute to basal cell cancer. Basal and squamous cell cancers often are cured if treated promptly. Melanomas are considerably rarer, have a poorer cure rate, and show a poorer correlation with UV exposure. Whether commonly used electric light sources provide enough UV radiation to increase carcinogenic risk is not certain. The unfiltered, quartz-bulb halogen lamps can emit enough UV radiation to induce actinic erythema in people who work under them for extended periods at high illuminances. Quartz halogen luminaires commonly include glass filters to reduce UV emissions. The Commission Internationale de l'Éclairage (CIE) concludes that there is insufficient evidence to support the hypothesis that common fluorescent lamps can cause malignent melanoma.53

Effects of Light on Vitamin D and Calcium Metabolism57-61 UV radiation plays an important role in the production of vitamin D in the skin (Figure 5-10). This vitamin is essential for normal intestinal absorption of calcium and phosphorus from the diet and for the normal mineralization of bone. Vitamin D deficiency causes a deficiency of calcium and phosphorus in the bones (such that they bend, fracture, or become painful) and causes such bone-softening diseases as rickets in children and osteomalacia in adults. Vitamin D poisoning, on the other hand, leads to excessive absorption of calcium and phosphorus from the diet and consequently a toxic effect on the skeleton. There is also a resultant increase in the blood calcium concentration and a precipitation of calcium phosphate deposits in vital organs, causing permanent damage or even death. Vitamin D poisoning also causes increased excretion of calcium in the urine, which can produce kidney stones or bladder stones. Mild cases of vitamin D poisoning lead only to increased urinary calcium excretion. Studies in animals and humans clearly show that UV-B radiation is effective in producing vitamin D in the skin. The action spectrum for this effect has been determined directly in human skin, with a peak of effectiveness near 297 nm. Melanin content in the skin, sunscreen use, and aging decrease the capacity of the skin to produce vitamin D. Furthermore, such environmental factors as changes in latitude, season, and time of day also greatly influence the cutaneous production of vitamin D. Increased exposure to sunlight results in an increased production of vitamin D, which can be detected in the blood. Most of the vitamin D requirement (upwards of 90%) for children and adults comes from casual exposure to sunlight. Elderly or infirm persons who consequently might not be exposed to normal environmental levels of UV radiation depend on dietary sources and supplements for their vitamin D requirement.

Light Versus Diet as Sources of Vitamin D. Natural foods contain little vitamin D, with the exception of certain fatty fish and fish liver oils. To reduce the dependence on environmental radiation, dairy products, cereals, and certain other foods sometimes are fortified with vitamin D. In some areas, the fortification of dairy products is accomplished with a form of vitamin D (D2) that is biologically similar to but chemically distinct from the vitamin D (D3) produced in the skin. Fortification has virtually eliminated childhood vitamin D deficiency in areas where it was formerly a common and serious public health problem, such as countries in high latitudes. Since many adults do not eat vitamin-D-fortified foods, they remain relatively dependent on cutaneous synthesis to meet their requirements for vitamin D. The current recommended dietary allowance for vitamin D in adults in the United States is 10 µg/day, or 400 international units. Excessive exposure to sunlight does not cause vitamin D intoxication; however, excessive vitamin D intake can occur, although this usually requires ingestion of over 1000 µg (40,000 international units) of vitamin D daily for some time.

Biological Rhythms6,8,62-99 Cyclic changes in biological parameters have been observed across species throughout the plant and animal kingdoms. These rhythmic alterations are loosely termed biological rhythms. Biological rhythms manifest themselves at both the macroscopic (multicellular) and microscopic (unicellular and subcellular) levels. Each rhythm has a characteristic amplitude, or magnitude of periodic change, and a characteristic period or frequency of oscillation. The timing of all biological rhythms involves the coordination, or entrainment, of external time cues (called exogenous zeitgebers) with an internal, or endogenous, pacemaker. External cues are for the most part derived from one or more of four geophysical cycles occurring in the natural environment: the tidal cycle, day-night cycle, lunar cycle, and seasonal cycle. Each natural cycle causes the synchronization of a particular rhythm (called a circarhythm). For example, circadian rhythms (entrained to the day-night cycle) are manifest in almost every plant (the raising and lowering of plant leaves throughout the day) and animal (the sleep-wake cycle). Seasonal rhythms, or circannual rhythms, are also widely manifest, as in plant seed germination and the seasonal breeding of many mammalian species. In addition to circarhythms, many cyclic patterns have been observed that cannot be directly linked to an external environmental synchronization. These rhythms tend to be shorter than one day (ultradian) as in the 90-min sleep cycle, or longer than one day (infradian), as in the female estrous or menstrual rhythm. In short, the timing of all biological rhythms, irrespective of period length, is dependent on both exogenous and endogenous entrainment cues. Circadian Rhythms. Rhythms that occur on an approximate 24-h schedule are termed circadian rhythms. Circadian rhythms are of particular interest because they characterize the pattern of variation observed in the majority of human physiologic rhythms, including body temperature, sleep pattern, hormone secretion, and blood pressure. Environmental light is the primary stimulus that mediates entrainment in the mammalian circadian system. Some evidence suggests that behavioral cues (social interaction) and artificial cues (alarm clocks) can also serve as entraining factors in humans. Evidence for an internal timing mechanism (or biological clock) can be produced by placing an organism under constant conditions (constant light or darkness), thereby denying it access to exogenous time cues. Organisms placed under such conditions continue to manifest circadian rhythms, but at a period characteristic of their own internal clocks. Such rhythms, which vary in period among species, are said to be free-running. In mammals, for example, nocturnal (nightactive) species tend to have faster internal clocks with periods less than 24 h, whereas diurnal (day-active) species tend to have slower clocks with periods greater than 24 h. The average human free-running period is longer than 24 h. Retinal and Ocular Physiology. In the mammalian circadian system, photic information is processed by the retina and relayed to the hypothalamus of the brain via a neural pathway called the retinohypothalamic tract (RHT). The retinal photoreceptors and photopigments employed in the visual system are discussed in Chapter 3, Vision and Perception. It remains to be determined which photoreceptors and photopigments are responsible for signal transduction in the circadian system. In examining studies from many laboratories, some data have indicated that the peak sensitivity of the circadian and neuroendocrine system is near 500 nm. This supports the hypothesis that rhodopsin or a rhodopsin-based molecule is the primary receptor for circadian and neuroendocrine regulation. Other data, however, have suggested that other photopigments might be involved in these regulatory effects. In rodents, for example, short wavelengths of sufficient intensity in the UV region of the spectrum as well as long wavelengths in the visible region are capable of suppressing melatonin (a hormone secreted by the pineal gland that follows a circadian pattern), entraining circadian rhythms, and influencing reproductive responses. Further studies are required to identify conclusively which specific photoreceptors and photopigments are involved in regulating the circadian and neuroendocrine systems among different animals. Ongoing studies are exploring the role of ocular elements for processing photic stimuli involved in circadian regulation. Specifically, the elements for stimulus processing of the circadian system include gaze behavior relative to the light source, the spectral characteristics of the light source, and the transmission of light through the pupil and ocular media. The ability of the circadian system to integrate photic stimuli spatially and temporally also is under study. In mammals, most investigators have operated from the assumption that retinal photoreceptive physiology in the eye, as

opposed to a photoreception in the skin or some other part of the body, is responsible for circadian and neuroendocrine regulation. Indeed, there are experimental data supporting this assumption in mammals.64,78,99 However, an intriguing theory proposed that bloodborne elements circulating through the eye might be responsible for transducing photic stimuli for circadian and neuroendocrine regulation.90 In one study, a single 3-hr bright light pulse of 13,000 lx delivered to the back of the knees of healthy humans systematically reset circadian body temperature and melatonin rhythms.71 In contrast, a similar bright light exposure failed to elicit acute melatonin supression in healthy humans.82 Further work is needed to determine whether or not the eyes are exclusive sites for circadian photoreception in humans.

Figure 5-11. Simplified illustration of pathway from the retina to the suprachiasmatic nucleus (SCN) or the hypothalamic "clock" and its long multisynaptic projection to the pineal gland by way of the paraventricular nucleus (PVN) in the hypothalamus. Note this pathway is anatomically separate from the pathway to the visual cortex, which serves the sensory capacity of vision. Neural Pathway. After retinal detection of photic information, the RHT projects directly to a bilateral structure known as the suprachiasmatic nucleus (SCN), which is believed to be the principal circadian pacemaker in mammals. Ablation of the SCN in rats and hamsters causes arrhythmia (loss of circadian rhythmicity). Neural projections from the SCN travel to many diverse control centers in the nervous system, including other areas of the hypothalamus as well as the thalamus, midbrain, brain stem, and spinal cord. One nerve pathway that carries nonvisual photic information extends from the SCN to the pineal gland (located in the brain itself) via a multisynaptic pathway, with connections being made sequentially in the paraventricular hypothalamus, the upper thoracic intermediolateral cell column, and the superior cervical ganglion (Figure 5-11). Cycles of light and darkness relayed by the retina entrain SCN neural activity, which in turn entrains the rhythmic production and secretion of melatonin from the pineal gland. In all vertebrate species studied to date, including humans, high levels of melatonin are secreted during the night and low levels are secreted during the day. In addition to entraining melatonin secretion from the pineal gland, light can have an acute suppressive effect on melatonin. Specifically, exposure of the eyes to light during the night can cause a rapid decrease in the high nocturnal synthesis and secretion of melatonin. Numerous studies have examined how the photic parameters of light intensity, wavelength, exposure duration, and timing interact with melatonin regulation. Human Biological Rhythms and the Circadian System. There are a myriad of measurable human physiological quantities that vary cyclically on a 24-h schedule.,8,63-65,73,78,80,83,84,86,87,97,98 The core body temperature reaches its peak during midday, dropping approximately 1°C to a nadir during sleep. Concentrations of various hormones (for example, melatonin) change regularly with the time of day. Other measurable quantities, including blood pressure, hand grip strength, alertness, cognitive performance, and visual sensitivity, also show circadian rhythmicity. The effects of drugs on the body vary with a circadian rhythmicity as well. Chronopharmacology examines rhythmic sensitivity to drugs. Certain drugs and dosages may elicit a response at one time of day but prove ineffective after another administration that same day. Sophisticated drug prescription and administration considers the circadian variation in drug response sensitivity.

Desynchronization of Rhythms. The various circadian rhythms detailed above are all synchronized by the internal clock, which is in turn entrained to the external 24-h daily schedule. Sleep deprivation, placement in constant conditions, and exposure to light during the night hours can cause a loss of entrainment to the environment. Thus, the internal clock can become phase shifted, or out of synchrony, with the external world. This desynchronization often contributes to feelings of discomfort, as commonly manifested in the jet lag syndrome. Symptoms of jet lag can include a disturbance in sleep and wakefulness, digestive difficulties, physical fatigue, menstrual irregularity, confusion and irritability, and reduced cognitive performance. Certain occupational requirements often involve mandatory phase shifting of the internal clock. Perhaps the most common of these is shift work. By the broadest definition, shift workers are persons who do not work a standard daytime schedule. Instead, they work nights, evenings, rotating shifts, split shifts, or extended shifts. In a report from the U.S. Congress,97 it was estimated that one out of five full-time workers in the United States is a shift worker. In agreement with many investigators, that report indicated that the two most common and destructive problems associated with shift work are a reduced quality of sleep following night work and a reduced capacity to maintain alertness while at work. These side-effects translate into increased accidents, decreased production, and performance deficits among those who are working at night. Furthermore, evidence indicates that shift workers have increased health problems, including a higher risk of cardiovascular disease, gastrointestinal distress, and cognitive and emotional problems. They also are likely to experience one or more of the symptoms attributed above to jet lag. In both conditions, chronic desynchronization of internal biological rhythms is thought to be a cause of such symptoms. Additional research is underway to study whether a link exists between lighting in the shift work environment,70 and melatonin deficiency or irregularity with respect to a risk of breast cancer. In addition to the currently understood effects of light on melatonin concentrations, the "melatonin hypothesis"95 suggests a link between melatonin disruption and the incidence of breast cancer. This is presently a controversial area of study and others have advanced the theory that there is no health risk associated with light, magnetic fields, melatonin, and breast cancer. Researchers believe that poor chronobiological adjustment to a permanent or rotating schedule causes some of these ailments. Not all of them, however, are solely due to a maladapted biological clock. In addition to a desynchronized circadian system, shift workers generally tend to be chronically sleep-deprived and experience domestic stresses that are more or less independent of circadian adaptation. On the frontiers of shift-work research, some investigators are attempting to develop strategies of light stimulation in an effort to improve circadian entrainment and to enhance performance and alertness in night workers. This new form of light therapy is discussed below.

Phototherapy8,86,97,100-179 Phototherapy, or the use of light as the primary or supplementary source of treatment for a disorder, is an established and burgeoning field. Light has been used therapeutically in a wide variety of applications, including dermatology, photochemistry, psychiatry, and oncology. Some forms of treatment, such as photochemotherapy, are established and have been practiced for decades, while others, such as low-level laser therapy, remain experimental. Retinal Photocoagulation. A therapeutic effect of both incoherent and coherent (laser) radiation in the 400- to 1400-nm wavelength bands involves photocoagulation techniques used to repair retinal detachment. The original coagulation process, involving the welding of the detached retina to the sclera, was accomplished with incoherent white light from a xenon lamp coagulator. The lamp has been superseded in most ophthalmological clinics by ruby, argon, and diode laser coagulators. Today, the photocoagulation technique in ophthalmology has been applied to the treatment of diabetic retinopathy, age-related maculopathy, and many other pathologies involving the eye. Phototherapy of Neonatal Hyperbilirubinemia.100-106 Hyperbilirubinemia in neonates is more commonly known as jaundice of the newborn. It is estimated that 60% of all infants born in the United States develop jaundice during the first week of life and that about 7 to 10% of neonates have hyperbilirubinemia of sufficient severity to require medical attention. Jaundice is the symptom and not the disease. It results from the accumulation of a yellow pigment, bilirubin, as a result of the infant's inability to rid itself of bilirubin as rapidly as it is produced. Bilirubin is chemically a tetrapyrrole and is derived principally from the degradation of hemoglobin. At normal concentrations, bilirubin is transported in the blood by binding to albumin. When the bound bilirubin reaches the liver, it is conjugated from a lipophilic to a hydrophilic substance that can be excreted in the urine. Infants with hyperbilirubinemia lack the ability to bind and excrete bilirubin in the normal manner.

In neonates, increased amounts of unconjugated bilirubin circulate in the blood. This is a result of normal red corpuscle degradation coupled with the functional immaturity of the neonatal liver. Peak levels of bilirubin between 5 and 13 mg/dL typically occur in healthy full-term neonates between the second and fifth day of life. By the seventh day of life, they typically decrease to normal adult levels. In the case of premature infants, bilirubin levels build up more slowly to reach peak levels betwen 10 and 15 mg/dL, and then slowly decline to adult levels over a period of up to four weeks. As the plasma concentration of bilirubin increases, there is a danger of exceeding the body's albumin-binding capacity, allowing free bilirubin to circulate. If unconjugated bilirubin reaches high levels (10 to 15 mg/dL) in a newborn, the pigment can penetrate the blood-brain barrier and accumulate in the brain, thus producing bilirubin encephalopathy and irreversible damage from toxic injury to brain cells, a condition known as kernicterus. Kernicterus often leads to the development of neurological injuries, including learning impairment, cerebral palsy, deafness, and in extreme cases death. After detection of hyperbilirubinemia, the condition can be monitored by measurement of the blood plasma bilirubin level. Phototherapy102-106 can be used to prevent the dangerous rise in plasma bilirubin. Typically, phototherapy is administered with one of three types of systems: a conventional or overhead system of fluorescent lamps, an overhead tungsten-halogen spotlight, or a relatively newer fiberoptic pad. The light sources may be filtered to maximize radiation in the short visible wavelength region and to minimize unnecessary UV and IR radiation. Overhead systems may be portable or incorporated into incubators, radiant warmers, or bassinets. They typically are mounted 25 to 50 cm from the infant, depending on the intensity required. Because of the blue appearance of the illumination from these systems, changes in infant skin color can be difficult to detect. Blue illumination also may contribute to irritation or nausea in some caregivers. For these reasons, the American Academy of Pediatrics (AAP) recommends a mixture of blue and white lamps in overhead phototherapy systems. During phototherapy, the infants are naked except for eye patches or goggles that protect the eyes from injury. In practice, such patches can be difficult to work with; they can slip off the eyes of particularly active infants and can interfere with circadian rhythms and parent-infant contact. Fiberoptic phototherapy pads developed in the late 1980s obviate some of the problems associated with overhead systems. The light from a tungsten-halogen lamp is delivered via fiberoptic cables to the pad, where they emit the light through the sides and ends of the fibers. Some pads can be wrapped around infants. If properly secured and covered, eye patches need not be used with phototherapy pads. The AAP suggests a minimum average spectral irradiance of 4 mW/cm2/nm in the range between 425 and 475 nm for photochemical reduction of bilirubin. Most medical textbooks recommend average treatment levels between 6 and 12 mW/cm2/nm. When bilirubin levels are dangerously high, many physicians use even higher irradiances, accomplished by moving sources closer to the infants in overhead systems, multiple overhead systems, or a combination of overhead and fiberoptic pad systems. Brief exposure to sunlight also can provide sufficient irradiance for phototherapy in full-term infants, provided that precautions are taken to avoid injury to the eyes and skin. The AAP recommends that phototherapy for full-term infants with nonpathological hyperbilirubinemia (no hemolysis) not begin until bilirubin levels reach concentrations of 15 to 20 mg/dL. If levels reach 20 to 25 mg/dL, exchange transfusions might be required, depending on the age of the infant and the judgment of the clinician. Phototherapy typically lasts for three days with preterm infants and for one to two days for full-term infants. Sometimes, phototherapy is performed in the home. The effect of phototherapy on plasma-free bilirubin concentration is shown in Figure 5-12. In contrast to exchange transfusion, phototherapy is noninvasive and poses less risk to the infant. Nevertheless, several side-effects have been observed (Figure 5-13). Phototherapy should be carried out only under the supervision of a suitably trained clinician. Phototherapy of Skin Disease.107-109 UV radiation is used for the treatment of various skin diseases such as psoriasis and eczema. The most effective wavelengths appear to be in the UV-B portion of the spectrum. Patients are usually given a small, whole-body exposure to a suberythemogenic or minimally erythemogenic dose of radiation three to five times a week. Usually twenty to forty such treatments are required to clear the skin. Maintenance treatments are then necessary at weekly intervals to control the condition until remission occurs. Various sources of radiation have been used, but at this time fluorescent and metal halide lamps are preferred. Adverse effects from this treatment are uncommon except for the short-term problem of erythema. Photoaging of the skin and presumably skin cancer are potential long-term problems, although the degree of risk of the latter effect has not been evaluated fully.

Figure 5-12. The effect of phototherapy of neonatal hyperbilirubinemia upon the mean serum bilirubin concentrations of 32 infants compared with that of 33 hyperbilirubinemic infants who received no treatment.

Figure 5-13. Side Effects of Phototherapy Photochemotherapy of Skin Disorders.117-122 Photochemotherapy is defined as the combination of nonionizing electromagnetic radiation and a drug to bring about a beneficial effect. Usually, in the doses used, neither the drug alone nor the radiation alone has any significant biologic activity; it is only the combination of drug and radiation that is therapeutic. PUVA (psoralen and UV-A) is a term used to describe oral administration of psoralen and subsequent exposure to UV-A. PUVA has proven to be effective in treating psoriasis, vitiligo, certain forms of severe eczema, a malignant disorder called mycosis fungoides, and a growing list of other skin disorders. Psoralens are naturally occurring tricyclic, furocoumarin-like chemicals, some of which can be photoactivated by UV-A. In living cell systems, absorption of energy from photons within the 320- to 400-nm waveband (with a broad peak at 340

to 360 nm) results in thymine-psoralen photoproducts and the transient inhibition of DNA synthesis. When certain psoralens are delivered to the skin either by direct application or by oral route, subsequent exposure to UV-A can result in redness and tanning, which are delayed in onset, occurring hours to days after exposure. The redness, or skin inflammation, from PUVA can be severe and is the limiting factor during treatment. The occurrence and degree of redness, however, is predictable and related to the doses of both the drug and UV-A irradiance. The redness that results from PUVA differs from sunburn in its time course. PUVA redness can be absent or just beginning at 12 to 24 h after UV exposure (when sunburn redness is normally at its peak) and peaks at 48 to 72 h or later. Because skin diseases can be treated at PUVA dose exposures that are less than those causing severe redness, careful dosimetry permits safe PUVA treatments. The pigmentation that results from PUVA appears histologically and morphologically similar to true melanogenesis (delayed tanning). Pigmentation reaches a maximum at approximately 5 to 10 days after PUVA exposure and lasts weeks to months. Psoriasis is a genetically determined hyperproliferative epidermal disorder. Until its cause or basic mediators are known, the most effective therapeutic agents must be those that have cytotoxic effects. Many such agents are effective but have potential cytotoxic effects on other than cutaneous organ systems. Since PUVA effects require UV-A, which penetrates into the skin but does not reach internal organs, PUVA offers the potential for combining the ease of systemic administration with the relative safety of limiting the biologic effects to the irradiated skin. Repeated PUVA exposures cause the disappearance of lesions of psoriasis in most patients. Ten to thirty treatments, given twice weekly, are usually adequate to achieve clearing. Weekly maintenance treatments keep most psoriatics free of symptoms. Psoriasis sometimes recurs weeks to months after PUVA therapy ceases. These patients respond to repeated PUVA therapy. The scalp, body folds, and other areas not exposed to UV-A do not respond to the therapy. During therapy, patients are typically exposed to UV-A 2 hours after ingestion of 0.6 mg/kg body weight of 8methoxypsoralen. The initial UV-A exposure (1.0 to 5.0 J/cm2) depends on the degree of melanization and on the sunburn history. The exposure must be increased as tanning occurs, because the pigmented skin diminishes UV-A penetration to the deeper levels of skin. Ideal radiation sources are those that have high radiant output of UV-A, the capability to irradiate the entire body surface, little UV-B and IR output, and uniform irradiance at all sites within the radiation chamber. Safety devices and reliable methods of measuring and delivering exact exposures are essential. The sun can be used as a PUVA radiation source but carries the disadvantage of unpredictable and varying UV irradiance and spectral distribution at the earth's surface. In tanned or pigmented patients, long exposure times can be required. For example, the exposure duration for both front and back of the body can be two to three times that needed for a single total-body treatment in a photochemotherapy system. Some patients, however, are willing to tolerate the heat and boredom of sun exposure in order to have the advantage of home treatment. Intense sun, clear skies, metering devices, careful instruction, and intelligent, cooperative, and motivated patients are required to make sun PUVA therapy a reasonable alternative to hospital or office treatment. Exposure to high irradiances of UV-A for prolonged periods of time can cause cataract and skin cancer in laboratory animals. These effects are enhanced by psoralens. The exposures used in these studies are much greater than therapeutic exposures. Observations in animal systems indicate that the extent of skin cancer induction varies with dose and route of psoralen administration and UV exposure. Both basal cell and squamous cell carcinomas have been observed in patients treated with PUVA. The incidence of these tumors is highest in patients with a prior history of exposure to ionizing radiation or a previous cutaneous carcinoma. These findings suggest that the potential risk of PUVA-related cutaneous carcinogenesis should be carefully weighed against the potential benefit of this therapy. Special care must be taken in treating patients with prior histories of cutaneous carcinoma or exposure to ionizing radiation. Experimental animal studies indicate that 8-methoxypsoralen also sensitizes the cornea and lens of certain species to UVA exposure. It is not yet known how this sensitization relates to the use of psoralens in photochemotherapy of humans. Although humans have used 8-methoxypsoralen therapeutically for decades, no cataracts attributable to PUVA have been reported. It seems wise, however, to limit the use of psoralen photochemotherapy to those with significant skin disease and to use adequate UV-A eye protection during the course of therapy. After ingesting psoralens, patients should protect their eyes for at least the remainder of that day. Physicians must be aware of these theoretical concerns and must carefully observe patients for signs of accelerated actinic damage. Glasses that are opaque to UV-A decrease total UV-A exposure to the lens and should be worn on treatment days. Photochemotherapy of Tumors.123-127 The photochemotherapy of tumors (also known as photodynamic therapy) employs visible radiation of a particular wavelength band as a catalyst in a photodegradation reaction. The products of this reaction are cytotoxic and effectively destroy tumor cells. The chemical hematoporphyrin derivative (HpD), when introduced into the blood, locates and binds to tumor cells. Exposure of the tumor to radiation at 630 nm causes the production of singlet oxygen from its previously bound triplet state in HpD. Singlet oxygen is highly cytotoxic and

consequently causes tumor cell degradation. Filtered xenon and tungsten lamps can be used to treat cutaneous lesions. A pumped dye argon laser radiating at 630 nm, connected to an optical delivery system such as fiber optics, can be used with an endoscope or similar device to reach internal cavities. Photodynamic therapy has achieved partial or complete response in 85% of patients with lung, esophageal, bladder, ocular, head and neck, neurological, and gynecological tumors. Despite this success, treatment generally has been limited to cutaneous and subcutaneous tumors (including breast cancers, melanomas and basal cell carcinomas). The photoreactivity of HpD also can be employed in tumor localization and detection, as radiation of 400 nm causes HpD to fluoresce. HpD is not toxic in the absence of light; however, as the substance is retained in the skin, it can cause photosensitivity that may persist for 3 to 4 weeks after infusion. Light Therapy for Seasonal Affective Disorder (SAD).99,128-147 During the past decade, the specific condition of fall and winter depression, or seasonal affective disorder (SAD), has been formally described in the scientific literature and included in the latest edition of the American Psychiatric Association's diagnostic manual, DSM-IV-R. Independent studies in the United States and Europe suggest that winter depression is a widespread syndrome. A study of the frequency of SAD manifestation on the east coast of the United States estimated that SAD occurs in less than 2% of the population in Florida, but in New Hampshire nearly 10% of the population show symptoms during fall and winter. From this study, it has been projected that as many as 10 million Americans have SAD and possibly an additional 25 million are susceptibility to a milder, subclinical form of SAD. People affected with this malady experience a dramatic decrease in their physical energy and stamina during the fall and winter months. As days become shorter, persons with SAD often find it increasingly difficult to meet the routine demands at work and at home. In addition to this general decrease in energy, SAD sufferers experience emotional depression, feelings of hopelessness, and despair. Other symptoms of winter depression or SAD can include increased sleepiness and need for sleep, increased appetite (particularly for sweets and other carbohydrates), and a general desire to withdraw from society. Fortunately, daily light therapy has been found to effectively reduce symptoms in many patients. There are now numerous clinics across North America that offer light therapy for people who are afflicted with winter depression. Specific treatment protocols vary somewhat among clinics. In the earlier days of light treatment for SAD, a patient often was instructed to sit a specific distance from a light panel that provided an illuminance of 2500 lx to the face. The patient was told not to gaze steadily at the bright light, but rather to glance directly at the unit for a few seconds each minute over a two-hour period. During the therapy period, a patient read, watched television, worked at a computer or did other hand work. Response to this therapy often was noted after two to seven days of light treatment. Benefits continued as long as the treatment was repeated regularly throughout the months that the individual experienced winter depression. Considerable research has been directed at determining the optimum illuminance, exposure, and time of day for the light treatment of winter depression. Most studies using light boxes or work stations indicate that illuminances from 2500 to 10,000 lx produce strong therapeutic results in treating SAD. In determining the best dosage of light, the intensity and exposure duration must be considered together. To date, no genuine dose-response functions have been established for light therapy, but exposure durations ranging from 30 min to 6 h in single or split sessions have been tested. The strongest therapeutic responses have been documented with a 2500-lx exposure over 2 to 4 h and with a 10,000-lx exposure over 30 min. Considerable data suggest that morning light treatment is superior, but not all investigators agree on this point. Current evidence supports the hypothesis that light therapy works by way of an ocular pathway as opposed to a dermal or transdermal mechanism. Several studies have investigated the action spectrum for SAD light therapy. Ultimately, a thoroughly defined action spectrum can both guide the development of light treatment devices and yield important information about the photosensory mechanism responsible for the beneficial effects of light therapy. Currently, it is premature to predict which photopigments or photoreceptors mediate the antidepressant effects of light. A practical issue debated among researchers concerns the role of UV in light therapy. Most of the early studies on SAD therapy used fluorescent lamps that emitted white light containing a small portion of UV-A energy. Those early findings erroneously led to the suggestion that UV-A is necessary for successful therapy. The current literature, however, clearly shows that SAD symptoms can be reduced by lamps that emit little or no UV. Hence, UV radiation does not appear to be necessary for eliciting positive therapeutic results. Most of the clinical trials treating winter depression have employed white light emitted by commercially available lamps. The white light used for treating SAD can be provided by a range of lamp types, including incandescent and cool white fluorescent. There is an assortment of light devices specifically designed for the treatment of SAD. Light therapy instruments come in a variety of shapes and configurations, including workstations, head-mounted light visors, and automatic dawn simulators. These devices are configured to shorten therapeutic time, increase patient mobility or permit therapy during the sleep period. Light visors appear to therapeutically benefit SAD patients with substantially lower energy requirements than those emitted by light boxes and work stations. Because dose-response comparisons have not been performed among different lamp types and light devices, it is not possible to distinguish which, if any, type of light

is superior for treating depression. Light Therapy for Jet Lag, Shift Work, and Sustained Performance.97,128-160 As scientists have explored the physiology of the human biological clock under normal conditions, they have also examined how that clock functions or dysfunctions under more unusual situations. Jet lag is a condition that results from rapidly moving across time zones. Although the human biological clock adjusts within three to seven days after such an event, during the adjustment period many people experience uncomfortable symptoms, which may include sleep and wake disruptions, gastric distress, irritability, depression, and confusion. Such symptoms can pose serious problems for the business traveler and can diminish the enjoyment of a vacation for the leisure traveler. Some studies have tested the use of light exposure to prevent or ameliorate jet lag symptoms. Investigators are optimistic that light can be a useful tool for the immediate resetting of the traveler's biological clock and can help overcome some of the problems associated with jet travel. It is generally agreed that there currently are insufficient data for a set prescription on how to best use light for this malady.97,159 Shift work poses a problem analogous to that found in jet lag. Instead of rapidly flying to distant places, shiftworkers may just as suddenly change the time period that they are awake or asleep. These individuals are awake and working during the night and attempting to sleep during the daylight hours. Although some individuals prefer night work over day work and are well adapted to shift schedules, shift work often is associated with decreased production, performance deficits, and increased health complaints. Some investigators have tested strategies of light stimulation to enhance performance and alertness in night workers. Studying simulated shift work over a 2- to 5-day period, different groups of investigators have shown that night workers had better circadian adaptation and improved alertness and cognitive performance when they worked under bright light (1,000 to 12,000 lx) than under dim light (100 to 150 lx). Other studies on simulated shift work have shown that exposure to bright white fluorescent light at specific times can improve sleep quality, enhance performance, and speed the adjustment of the circadian system. The U.S. military currently has a triservice research program aimed at finding ways of enhancing physical and mental performance in personnel who are on continuous duty for prolonged periods of time. A major focus of this program has been to study pharmacological agents that can help sustain alertness without degrading performance. Sustained performance studies have shown that workers exposed to bright light (3000 or 5000 lx) exhibited significantly improved behavioral and cognitive performance on selected tasks compared to their own performance on a separate occasion under 100 lx. In addition to these behavioral effects, there were significant differences in the body temperatures and plasma melatonin levels associated with light stimuli. In these acute studies, it is not clear how light influences performance. There is, nevertheless, a consensus among scientists that it is still premature to formulate a prescription on how to best use light for both short-term and long-term work applications.97,159 Potential Placebo Responses in Light Therapy, Mood, and Performance Effects.161,162 In considering the newer uses of light for therapeutically reducing the symptoms of winter depression, jet lag, and shift work, it is important to examine whether the observed effects are due to specific light therapy or to a nonspecific or placebo response. When using light experimentally on humans, either for therapeutic purposes or for work or travel applications, investigators are confronted with a dilemma. Simply put, when volunteers can readily see that a manipulation of light is part of an experiment, there is a distinct possibility of finding a placebo reaction to the light treatments. In the medical literature it has been well documented that patients with a wide range of disorders--depression, schizophrenia, and anxiety as well as cancer, diabetes, and ulcers--can respond to placebo treatments. Hence it is likely that SAD patients, world travelers, and shift workers show some level of placebo response to light therapy. The degree to which the patients' response to light therapy is due to a nonspecific placebo response or to a genuine clinical response remains an open question. Low-Level Laser Therapy.163,164 Although not yet approved for routine medical use in the United States, low-level lasers at 633, 830, and 904 nm are used widely throughout the world in sports medicine clinics and by veterinarians to accelerate wound healing, treat sprains, and control certain types of pain. Unfortunately, low-level lasers also are used to treat other conditions for which there is little hard evidence of benefit. The scientific community should encourage this fledgling field to establish proper controls and to learn more photobiology in order to establish unequivocally which clinical conditions are improved by this type of therapy and which are not. Attempts have been made to explain the photobiological basis of how visible and IR radiation can produce similar clinical and cellular responses. Photorefractive Keratectomy. A surgical technique known as photorefractive keratectomy (PRK), using lasers to sculpt the cornea of the eye, thus altering its refractive power, has been developed and tested in Europe and the United States. It has been used to correct refractive errors such as myopia and astigmatism. The use and acceptance of this technique appears to be increasing; however, clinical trials to assess long-term effects are currently underway.165 Lighting Safety Criteria.166-175 Human exposure limits for nonionizing optical radiation are consensus values. The Threshold Limit Values (TLVs) of the American Conference of Governmental Industrial Hygienists (ACGIH)171

normally are used in the United States and are widely accepted internationally. These TLVs are reviewed and updated annually to represent the best current scientific consensus for exposure safety. It is explicitly stated that these TLVs "represent conditions under which it is believed that nearly all workers may be repeatedly exposed without adverse health effects." Because they are presented as specific values, concern might arise if an exposure exceeds one of these values. The ACGIH explicitly addresses this concern by stating that the TLVs are guidelines, not specific breakpoints between safe and dangerous exposures. The TLVs are the basis of the ANSI/IESNA RP-27.1-96 recommended practice.173 This document covers optical radiation of lamps and lamp systems between 200 nm and 3000 nm except for lasers and light-emitting diodes used in optical fiber communications. It expands upon and details methods for applying TLV criteria, which can be described as follows: 1. UV actinic effects of photokeratitis and photoconjunctivitis of the eye, and erythema (sunburn) of the skin. A spectral weighting function from 200 to 400 nm is used to collectively represent the potential hazard of radiation with respect to these effects.171 2. UV cataractogenesis. Until the possibility of an increased risk of cataracts owing to long-term exposure is resolved, ocular exposure to radiation between 320 and 400 nm should be limited as a precaution. 3. Retinal photochemical injury ("blue-light" hazard). The retinal image of a source with high levels of energy primarily between 400 and 500 nm can produce photochemical injury of the retina. Radiation between 400 and 700 nm is spectrally weighted by a function to establish the potential for injury. 4. Retinal thermal energy. Viewing a high-radiance source can elevate retinal temperature. The radiant power between 400 and 1400 nm is spectrally weighted by a function related to ocular transmittance and retinal absorbance. Because retinal heat transfer depends on the image area, this criterion includes the angular size and shape of the source. This type of injury is dominant over retinal photochemical injury for exposures less than 10 s. 5. IR cataractogenesis. Chronic exposure to high levels of irradiance between 770 and 3000 nm can increase the risk of certain types of cataracts.168 6. Skin thermal injury. Cellular injury occurs if skin temperature reaches approximately 45°C. Because this temperature is associated with intolerable pain, injurious exposure tends to be self-limited by discomfort for extended exposure times, and this criterion is applied only to short duration exposure to radiation between 400 and 3000 nm. These criteria are applied to specific exposure situations. Another recommended practice174 extends these criteria to develop risk group classification for lamps. Lamps are divided into four groups each associated with a degree of potential hazard. The absolute degree of risk or safety cannot be determined for most lamps independent of their specific use in an application. The four risk groups and the philosophical basis for each of them are as follows: 1. Exempt group: The lamp does not pose any photobiological hazard within the limits specified in ANSI/IESNA RP-27.3.174 2. Risk group 1 (low risk): The lamp does not pose any photobiological hazard due to normal behavioral limitations on exposure. 3. Risk group 2 (moderate risk): The lamp does not pose any photobiological hazard due to the aversion response to very bright sources or due to thermal discomfort. 4. Risk group 3 (high risk): The lamp may pose a photobiological hazard even for momentary or brief exposures. ANSI/IESNA RP-27.3174 defines exposure conditions (including time and distance) based on the philosophy of the risk groups. Using the characteristics of a lamp, the resulting exposures are evaluated in accordance with the criteria of ANSI/IESNA RP-27.1173 to determine the risk group classification for the lamp. The system places a lamp in a single risk group based on the likelihood and seriousness of the potential risk. Specific lamp labeling and informational requirements are specified for each risk group. Owing to concern about eye safety and products that incorporate laser-type emitting devices, including certain lightemitting diodes (LEDs, see Chapter 6, Light Sources), the International Electrotechnical Commission (IEC) and

European Committee for Electrotechnical Standardization (CENELEC) have developed standards to minimize risks of eye injury from use of products containing LEDs. These standards include maximum permissible exposure (MPE) levels and required testing methods for products using LEDs, as well as eye safety labeling recommendations based on the amount and type of emission produced by these products, just as with other light sources.168

EFFECTS ON MICROORGANISMS Germicidal (Bactericidal) UV Radiation176-191 Electromagnetic radiation in the wavelength range between 180 and 700 nm is capable of killing many species of bacteria, molds, yeasts, and viruses. The germicidal effectiveness of the different wavelength regions can vary by several orders of magnitude, but wavelengths shorter than 300 nm are generally the most effective for bactericidal purposes.

Figure 5-14. Tentative Germicidal (Bactericidal) Efficiency of Ultraviolet Radiation at Mercury Emission Lines The bacterium most widely used for the study of bactericidal effects is Escherichia coli. Studies have shown the most effective wavelength range to be between 220 and 300 nm, corresponding to the peak of photic absorption by bacterial deoxyribonucleic acid (DNA). The absorption of the UV radiation by the DNA molecule produces mutations or cell death. The relative effectiveness of different wavelengths of radiation in killing a common strain of E. coli is shown in Figure 5-14. Germicidal (Bactericidal) Lamps. The most practical method of generating germicidal radiation is by passage of an electric discharge through low-pressure mercury vapor enclosed in a special glass tube that transmits shortwave UV radiation. Approximately 95% of the energy from such a device is radiated at 253.7 nm, which is very close to the wavelength corresponding to the greatest lethal effectiveness. These lamps come in various sizes and shapes including linear and compact sources. Hot-cathode germicidal lamps are similar in physical dimensions and electrical characteristics to the standard preheat 8-, 15-, and 30-W fluorescent lamps. While both types of lamps operate on the same auxiliaries, germicidal lamps contain no phosphor and the envelope is made of a UV- transmitting glass. Quartz envelopes are used for some germicidal lamps. Slimline germicidal lamps are instant-start lamps capable of operating at several current densities within their design range, 120 to 420 mA, depending on the ballast with which they are used. Cold-cathode germicidal lamps are instant-start lamps with a cylindrical cathode. They are made in many sizes and operate from a transformer. The life of the hot-cathode and slimline germicidal lamps is governed by the electrode life and frequency of starts. (Their effective life is sometimes limited by the transmission of the bulb, particularly when operated at low temperatures.) The electrodes of cold-cathode lamps are not affected by the number of starts, and their useful life is determined entirely by the transmission of the bulb. All types of germicidal lamps experience a decrement in UV emission as the total hours of operation increase. Lamps should be checked periodically for UV output to ensure that their germicidal effectiveness is maintained. The majority of germicidal lamps operate most efficiently in still air at room temperature. For lamp efficiency measurements, UV output is standardized at an ambient temperature of 25°C. Temperatures either higher or lower than

this value decrease the output of the lamp. Slimline germicidal lamps operated at currents ranging from 300 to 420 mA and certain preheat germicidal lamps operated at 600 mA are designed exceptions to this general rule. At these high current loadings, the lamp temperature is above the normal value for optimum operation; therefore, cooling of the bulb does not have the same adverse effect as with other lamps. These lamps are well suited for use in air conditioning ducts. In addition to emissions at 253.7 nm, some germicidal lamps generate a controlled amount of 184.9-nm radiation, which produces ozone (Figure 5-15). Since ozone is highly toxic, its environmental concentrations have been limited by an Occupational Safety and Health Administration (OSHA) regulatory mandate to 0.1 parts per million (ppm), or 0.2 mg/m3. Care should be taken when choosing germicidal lamps to meet the requirements of these regulations.

Figure 5-15. Relative spectral distribution of energy emitted by ozone-producing germicidal lamps. Photoreactivation. It has been observed that the survival of UV-irradiated bacteria could be greatly enhanced if the cells were subsequently exposed to an intense source of blue light. Researchers have demonstrated the existence of a photoreactivating enzyme and established its basic properties in repair of damaged DNA. The enzyme combines in the dark with cyclobutyl pyrimidine dimers in UV-irradiated DNA to form an enzyme-substrate complex. When the complex is activated by the absorption of energy between 320 and 410 nm, the cyclobutyl pyrimidine dimers are converted to monomeric pyrimidines and the enzyme is released. Under certain experimental conditions, as much as 80% of the lethal damage induced in bacteria by low-energy UV radiation at 253.7 nm can be photoreactivated, thus indicating the importance of cyclobutyl pyrimidine dimers as lethal lesions. Photoreactivating enzymes have been found in a wide range of species, from the simplest living cells to the skin and white blood cells of humans. Germicidal Effectiveness. The effectiveness of germicidal radiation is dependent on many parameters, including the specific susceptibility of the organism, the wavelength of radiation emitted, the radiant flux, and the time of exposure. Figure 5-16 lists the exposure intensity (J/m2) of the 253.7-nm UV radiation necessary for the inhibition of colony formation (a 90% reduction in population) in a wide variety of microorganisms.

Figure 5-16. Incident Radiation at 253.7 nm Necessary to Inhibit Colony Formation in 90% of the Organisms

Figure 5-17. Germicidal lamps for air disinfection in occupied rooms: (a) open unit used in rooms over 2.7 m (9 ft) in height; (b) louvered unit used where ceilings are lower than 2.7 m (9 ft). Dimensions: A, 3.7 m (12 ft);

B, 2.1 m (7 ft); C, 2 m (6.5 ft); D, 2.7 m (9 ft). Germicidal effectiveness is proportional to the product of intensity and time (from 1 µs to several h). A nonlinear relationship exists between UV exposure and germicidal efficacy. For example, if a certain UV exposure kills 90% of a bacterial population, doubling the exposure time or intensity can kill only 90% of the residual 10%, for an overall germicidal efficacy of 99%. Likewise, a 50% decrease in intensity or exposure time decreases germicidal efficacy only from 99% to 90%. Humidity can reduce the effectiveness of germicidal UV radiation.

Figure 5-18. Reflectance of Various Materials for Energy of Wavelengths in the Region of 253.7 nm Precautions. Exposure to germicidal UV radiation can produce eye injury and skin erythema and has produced skin cancer in laboratory animals. The ACGIH limit for exposure of the unprotected skin or eyes to radiation at 253.7 nm is 6 mJ/cm2 within an 8-h period. For example, this conservative limitation would be 0.2 µW/cm2 for an 8-h continuous exposure, 0.4 µW/cm2 for a 4-h continuous exposure, and 10 µW/cm2 for a 10-min continuous exposure. The maximum exposure time is only 1 min for 100 µW/cm2. Some common G30T8 unshielded germicidal lamps can deliver this irradiance from 0.75 m. Based on the potential for producing threshold keratitis, the National Institute of Occupational Safety and Health (NIOSH) has proposed that half of the intensity-time relationship established by ACGIH above be used as a safe industrial exposure for the eye. Eye protection is essential for all who are exposed to the direct or reflected radiation from lamps emitting UV radiation, especially those germicidal lamps emitting UV-C radiation. Ordinary window or plate glass or goggles that shield the eyes from wavelengths shorter than 340 nm are usually sufficient protection. However, if the radiation is intense or is viewed for some time, special goggles should be used. Failure to wear proper eye protection can result in temporary but painful inflammations of the conjunctiva, cornea, and iris; photophobia; blepharospasm; and ciliary neuralgia. Skin protection, achieved by wearing clothing and gloves that are opaque to germicidal radiation, is advised if the UV radiant intensity is high or if the exposure duration is long. Accidental overexposure can be avoided by education of maintenance workers. Warning signs in appropriate languages should be posted.

Applications Air Disinfection in Rooms. With the resurgance of multiple-drug-resistant forms of airborne disease (e.g., Mycobacterium tuberculosis), new attention is being given to using UV air-mixing systems to prevent transmission. These systems can provide cost-effective controls in strategically placed areas and possibly in the whole building. In occupied rooms, irradiation by an open-luminaire germicidal lamp should be confined to the area above the heads of occupants as shown in Figure 5-17. The ceiling of the room to be disinfected should be higher than 2.9 m (9.5 ft), and occupants should not remain in the room for more than 8 h. If either of the above conditions does not meet the requirements of the workspace, louvered equipment should be used to avoid localized high concentrations of flux that may be reflected onto room occupants. Louvered luminaires using compact sources and electronic ballasts can provide energy efficient wall-, corner-, and pendant-mounted upper-room options. Some of these luminaires meet OSHA and NIOSH limits for rooms with 2.9 m ceilings for surface-mounted units and pendant units at a height of at least 3 m. An average irradiation of 20 to 25 µW/cm2 is effective for slow circulation of upper air and maintains freedom from respiratory disease organisms comparable to outdoor air. Upper-air disinfection, as practiced in such areas as hospitals, schools, clinics, jails, shelters, transportation systems, and offices, can be effective in providing relatively bacteria-free air at the breathing level of room occupants. Personnel movement, body heat, and winter heating methods create convection currents through a room sufficient to mix upper and

lower room air. All surfaces irradiated by UV germicidal radiation (including ceilings and upper walls) should have a UV reflectance below 5% (characteristic of most oil and some waterbase paints) (Figures 5-18 and 5-19). "White coat" plaster or gypsum-product surfaced wallboard and acoustical tile can have higher germicidal reflectances and should always be painted with a less reflective substance. Unpainted white plaster walls and ceilings can limit safe exposure to only 2 to 3 h even with louvered luminaires. These precautions are especially important in hospital infant wards because children can be more sensitive to UV radiation than adults. In operating rooms where prolonged surgery is performed, UV sources are mounted above doorways to disinfect air entering through the doorways. Face and skin protection are required for anyone passing through these doorways. Air Duct Installations. It is possible to provide a sufficiently high level of UV radiation to kill 90 to 99% of most bacteria within very short exposure times at usual duct air velocities. Duct installations are especially valuable where central air heating and ventilating systems recirculate air through all of the otherwise isolated areas of a building. Slimline germicidal lamps, especially designed for cool, high-velocity ducts, commonly are installed inside access doors in the sides of ducts, either along or across the duct axis. Where possible, the best placement for lamps is across the duct to secure longer travel of the energy before absorption by the duct walls and to promote turbulence to offset the variation in UV radiation levels throughout the duct. Lamps should be cleaned periodically because dust buildup lowers UV emission. Sanitization Techniques. Three general methods of germicidal lamp placement can be employed to establish a sanitary environment: upper-air irradiation, barrier-type irradiation, and direct irradiation. As previously outlined, upper-air irradiation maintains purified air at the normal breathing level of room occupants. It also permits safe continuous occupancy. Barrier irradiation techniques employ a narrow beam of UV directed across an opening, effectively preventing live organisms from passing from one space to another. Direct irradiation exposes whole surfaces to germicidal radiation. The most effective and efficient form of sanitization, direct irradiation, is hazardous to room occupants. In such conditions, proper eye and skin protection must be worn while germicidal lamps are operated. Alternatively, direct irradiation can be used when the room is unoccupied; provisions to prevent entry during irradiation are therefore unnecessary.

Figure 5-19. The spectral reflectance of (a) indoor, (b) outdoor, and (c) natural materials. Liquid Disinfection. UV disinfection of water is used when it is essential to eliminate residual substances or taste. UV radiation is absorbed by natural chemical contaminants in water (e.g., iron-based and organic compounds) and by the DNA of water-dwelling organisms. Hence, disinfection of water often involves exposure times and intensities 40 to 50 times greater than those used in air sanitation. Such exposures are secured by slow gravity flow of water through shallow tanks under many lamps, or by immersing lamps enclosed in quartz tubes directly into the water. Liquids of high absorbance (e.g., fruit juices, milk, blood, serums, and vaccines) are disinfected with various film spreaders. These range from high-speed centrifugal devices and surface-adhering rotating cylinders to gravity-flow-down screens and inclined planes. Such devices spread a film of liquid to approximately the thickness of its molecular size. Disinfection of Granular Material. The surfaces of granular materials (e.g., sugar) are disinfected on traveling belts of vibrating conveyors designed to agitate the material during travel under banks of closely spaced germicidal lamps. In the case of sugar, thermoduric bacteria survive the vacuum evaporator temperatures of a sugar-syrup concentration and, forced out of the sugar crystals during lattice formation, remain in the final film of dilute syrup left on the crystal surface. Ordinarily harmless, these bacteria can cause serious spoilage in canned foods and beverages.

Product Protection and Sanitation. Product protection and sanitation are achieved by both air disinfection and surface irradiation (as with granular material). In this field, however, the usefulness of germicidal UV radiation generally is limited to the prevention of contamination during processing rather than the disinfection of product. For example, where sufficient irradiation to kill mold spores might be impractical, the vegetative growth of mold can be prevented by continuous irradiation at levels lethal to ordinary bacteria. Germicidal lamps installed in concentrating reflectors are used to disinfect air that might contaminate a product during processing and packaging, as in the travel of bottles from washing to filling to capping. Lamps serve to replace or supplement heat in processes where sterilization by heat might be destructive. Intensive irradiation of container surfaces can also supplement or replace washing between uses. UV sanitization techniques are used in bakeries, breweries, and packaging plants for liquid sugar, syrup, fruit juices, and beverages.

EFFECTS ON INSECTS Insect Responses192-199 The increasing popularity of outdoor living, drive-in businesses, and outdoor recreational establishments is accompanied by intensified insect problems caused particularly by nocturnal insects attracted to light. Similar problems are encountered at lighted farmsteads, animal pens, feedlots, processing plants, and industries operating at night in lighted facilities. Many of these problems can be prevented or greatly reduced if the responses of the insect pests are considered when designing and planning these facilities. The insect nuisance problems associated with lighting have four distinct but related aspects: the existing insect population in the surrounding vicinity, the attractiveness to insects of the activity carried on in the lighted area, the attractiveness of the lighting system used, and the suitability of the area for sustaining insect life. The circumstances of each situation are different, and usually little can be done about the insect population in the vicinity. A knowledge of the insects in relation to their normal habitats and of the activities to be carried on in the desired insectfree areas usually helps to anticipate problems. Preventive action can be based on the known behavioral patterns of the expected insects. Insects likely to cause problems can be broadly categorized as follows: Insects Not Attracted by Light. Insects not necessarily attracted to light are indigenous to enclosures and buildings, and most live continuously within the area. Included are cockroaches, ants, and flour beetles. Diurnal Insects. These insects normally live outside the area and are active during the day. They are attracted to the area for food, shelter, or breeding sites. Examples are houseflies, pomance or fruit flies, and honey bees. Nocturnal Insects. This group usually feeds and lives beyond the area and is attracted by electromagnetic radiation. These insects are the true problem associated with lighting systems. Most nocturnal insects, including moths, leafhoppers, mayflies, caddisflies, and various beetles, midges, and mosquitoes, are capable of flight and are active primarily at night. Phototaxis is the term applied to the movement of insects toward electromagnetic radiation. Insects that are attracted to a radiation source are said to be photopositive or to exhibit positive phototaxis. The spectral region most attractive to a wide range of insect species, especially nocturnal species, is in the UV-A range. Other species are known to respond to energy in the visible and IR regions as well. Research and experience have shown that insect problems can be greatly reduced with properly designed lighting systems and with proper management of the area.

Lighting System Attraction One means of reducing the insect nuisance is to select light sources having low insect attractiveness. In practice, this involves sources with high energy at long wavelengths (yellow or red) and low energy in the UV and short visible wavelengths (blue). High-pressure sodium lamps have approximately one-third as much associated insect nuisance as a comparable-wattage mercury lamp system. Incandescent and warm fluorescent lamps designed to minimize short visible wavelength and UV radiation are available. If insect-attractive lamps must be used, they can be shielded so that all their radiant output is confined to the area to be illuminated. If lamps must remain visible from the outside, consideration should be given to the use of refractors, filters, and shields made of glass or plastic material to filter out the UV radiation. If lamps emitting short-wavelength (blue) light and near-UV energy are used for lighting, they should not be directly visible at distances beyond a few meters from the illuminated area. Exposed lamps, including incandescent, should not be located directly over entrances because insects are attracted to these bright sources and gain access to the interior when a door is opened. Any type of lamp used to light an entrance or work area should be located a short distance away, with the light directed toward the area to be illuminated. If a lamp must be placed near the entrance, it should be shielded so that its radiant output is directed downward and confined to the

immediate area. Since insects are also attracted to reflected radiant energy, care should be taken to avoid using a surface or a paint with a high reflectance for short visible wavelength or UV radiation.

Decoy Lamps and Insect Traps In addition to careful lamp selection and shielding designs, the number of night-flying insects within an area also can be reduced by placing attracting "black light" fluorescent or mercury decoy lamps at 30- to 60-m distances around the perimeter of the area to intercept those insects trying to enter. Insect traps commonly contain black light lamps to attract photopositive insects as a means of killing or trapping them. One of the common killing mechanisms is an electric grid that electrocutes the insects attracted to the black light. Various designs are available for commercial, industrial, and residential use. The placement and number of traps should vary with the individual situation and the species of insects involved. Specialists in this field usually are required to determine the best placement of traps for solving specific insect problems. If grid traps are used, they should not be placed so that electrocuted insects fall or are blown into working and food processing areas. System designs and installations should be in compliance with the National Electrical Code or the Canadian Electrical Code.199 In agriculture, various insect traps have been used for survey purposes to detect insects in a crop area, to predict the need for pesticide application, and to evaluate effects of insecticide measures. Survey traps used over large areas can be used to determine migration of insects and to predict potential infestation. The trap designs usually include a black light or other fluorescent lamp and a means of trapping insects. Designs are altered for specific insect species. Studies with black light insect traps in large tobacco- or tomato-growing areas of over 300 km2 (10 mi2) have shown that one or more traps per square kilometer reduce tobacco and tomato hornworm populations. Indications are that insect traps are best used in conjunction with insecticides. The use of such traps would generally reduce the number of insecticide applications in a growing season.

EFFECTS ON POULTRY200 Egg Production Light affects egg production of laying chickens by causing the release of LH and FSH hormones from the anterior pituitary gland. These hormones increase the growth of ova in the ovary. A minimum daylength of 11 to 12 h is required for the effect to take place. Optimal daylength is 14 h, and there are indications that daylengths in excess of 17 h can decrease egg production. The most effective wavelengths are 664 to 740 nm, the region of peak spectral response of the chicken's visual system. The hormone secretions are governed only by light received by the eye and not by other areas of the body. A minimum illuminance of 10 lx at bird level is generally recommended for egg production. Additional illuminance gives no additional benefit (Figure 5-20a) and can decrease productivity by increasing such behaviors as hyperactivity, pecking, and cannibalism. This concern applies equally to the production of meat chickens.

Chick Growth and Development Light also affects the development of growing chickens (Figure 5-20b). As with the productivity of layers, daylength is the governing factor rather than illuminance, provided a threshold of approximately 10 lx is maintained. When daylength is reduced below 11 to 12 h during the growing period, time to sexual maturity is delayed by up to 3 weeks. The size of eggs during the first production period is increased without significantly affecting the total number of eggs produced. This increases profit because larger eggs are usually in greater demand. Consequently, daylight is typically excluded from commercial laying houses and illumination programs are controlled strictly. See Chapter 27, Lighting Controls.

Figure 5-20a. The Effect of Illuminance on Total Egg Production During the First 45 Weeks of Lay

Figure 5-20b. The Effect of Light Treatment Upon Sexual Maturity and Number of Eggs Produced During the First 47 Weeks of Lay

Embryonic Growth Photoacceleration is a poorly understood but effective technique for improving embryonic growth. It involves placing the large end of the egg as close as possible to a white light source without overheating. The effects are summarized as follows:

Reduced embryonic mortality due to accelerated growth during the critical first hours of development Earlier hatching, by 20 to 48 h Increased weight at hatching, by up to 15% Increased weight thereafter, by up to 100 g Earlier sexual maturity

The process requires continual exposure during the first 18 days.

EFFECTS ON PLANTS201-224 Because plants and humans share the same terrestrial environment, daylight has been important to the evolution of their respective spectral sensitivities. Most of the energy incident on earth from the sun is in the visible portion of the electromagnetic spectrum (Figure 5-1). However, plants and the human visual system have evolved very different spectral sensitivities. Strictly speaking, the term "light" is reserved for humans as visually effective radiant energy. Nevertheless, "light" also is used loosely to describe all radiant energy within the visible portion of the electromagnetic spectrum; for the purposes of discussion in this section, the term "light" is used in this latter sense.

Figure 5-21. Relative quantum yield for photosynthesis (a), the action spectra for the short-wavelengthabsorbing photoreceptor (b), the phytochrome in its long-wavelength-induced form (c), and the infraredinduced form (d) compared to human photopic vision (e).

Plant Responses Photosynthesis. Plants respond to light in many ways (Figure 5-21). Light provides the energy necessary for the conversion of carbon dioxide and water by chlorophyll-containing plants into carbohydrates in the process known as photosynthesis. These carbohydrates are essential foods and are the substrate for proteins, fats, and vitamins required for the survival of all other living organisms. Oxygen, formed as a by-product, is the major source of atmospheric oxygen. Most fossil fuel resources are derived from photosynthetic processes of a past geological period. Light is also essential for the formation of such important plant pigments as chlorophyll, carotenoids, xanthophylls, anthocyanins, and phytochrome. Light is effective in the opening of stomates, the setting of internal biological clocks, and the modification of such factors as plant size and shape; leaf size, movement, shape, and color; internodal length; flower production, size and shape; petal movement; and fruit yield, size, shape, and color. Respiration. The reverse of photosynthesis is respiration, whereby the carbohydrates formed in photosynthesis are oxidized to carbon dioxide, water, and energy. Respiration requires neither light nor chlorophyll, but it does require food, enzymes, and oxygen. Respiration is continuous, whereas photosynthesis occurs only in light. At moderate to high irradiance, photosynthesis in plants exceeds respiration, so the net effect is the production of oxygen from leaves. If irradiance decreases to the point where the carbohydrate produced is equal to that used in respiration, the apparent photosynthetic rate is zero and there is no diffusion of gas from the leaf pores (stomata). This phenomenon is called the compensation point. Plants lighted at the compensation point cannot survive long, because stored carbohydrate is used during the dark period. When the carbohydrate reserves are depleted, the plant dies. This is an important fact in the maintenance of plants in interior environments. Many plant species under high light levels experience an increase in respiration. This phenomenon, called photorespiration, results in a decrease in the apparent photosynthetic rate. Other Photoresponses. Photomorphogenesis is light-controlled enlargement, development, and differentiation of a plant due to responses initiated by the short-wavelength-absorbing photoreceptor and by phytochrome. The spectral responses of the major photoreceptors are shown in comparison with that of human vision (Figure 5-21). The shortwavelength-absorbing photoreceptor appears to be a flavin but may differ among species of plants. Phytochrome is a blue-green biliprotein; it has a chromophore that absorbs radiant energy and undergoes excitation, which is used to change its molecular structure. The photomorphogenic responses include flowering, seed germination, stem elongation, and anthocyanin pigment formation.

Figure 5-22. Relative photosynthetic rate in relation to irradiance and carbon dioxide concentration. Light-induced movements of plants are phototropism, photonasty, and phototaxis. Phototropism is the bending of an organ toward or away from the direction of the source of light and is controlled by a photoreceptor that is responsive to short-wavelength radiation. Photonasty is the movement of plant organs, such as the closing of flowers at night and opening during the day, due to changes in irradiance. Phototaxis is the movement of the whole organism in response to light and is restricted to sex cells of aquatic plants and unicellular aquatic plants.

Limiting Factors for Growth In addition to light, which provides the energy for plants, other requirements must be available in optimum amounts for rapid photosynthesis and growth. These requirements are water, nutrients (inorganic salts), suitable temperature, and carbon dioxide. Lack of any one of these places the plant in stress and limits or halts growth. The relationship between carbon dioxide and irradiance is shown in Figure 5-22, where it can be seen that the photosynthetic rate is accelerated by an increase in irradiance or in the carbon dioxide level. These principles are important in the application of light to accelerate plant growth.

PLANT LIGHTING Plant lighting is the application of light sources for the control of growth, flowering, or maintenance of plants.

Light Sources Electric lamps that emit sufficient energy between 300 and 800 nm are effective in photosynthesis and other photoresponses of plants. Experimental work in plant science uses many types of light sources, including incandescent, fluorescent, xenon, low- and high-pressure sodium, and metal halide lamps. Various combinations of lamp types are sometimes used, the most common combinations being incandescent and fluorescent. Other combinations include highpressure sodium plus metal halide and metal halide plus incandescent. See Figure 5-23. Generally, the most efficacious lamps for plant growth provide the greatest portion of their energy between 580 and 700 nm. Of the various fluorescent lamps, common rare-earth lamps appear to be efficient in the synthesis of dry matter in plants. Incandescent lamps, which emit a greater proportion of long-wavelength radiation, are widely used in the control of flowering horticultural crops. Special fluorescent lamps, termed "plant growth lamps," have been developed with phosphors providing emissions that match the absorption maxima of chlorophyll (Figure 5-24). These lamps have found some use in residential lighting for the growth and color enhancement of house plants requiring low energy levels, especially African violets and gloxinias. They are more expensive and have not been found to be superior to common rare-earth fluorescent lamps for photosynthesis. They are therefore not in general use.

Figure 5-23. This modern greenhouse in a college science building has high pressure sodium lighting in addition to large expanses of glazing that allow for plant growth with daylight.

Radiant-Energy Measurement As previously noted, plants respond to radiant energy quite differently than the human eye (Figure 5-21). Therefore, it is not accurate to measure plant irradiance in terms of illuminance. This is especially important when comparing the effects on plants of lamps with different spectral power distributions. Photosynthesis is a photochemical conversion in which each molecule is activated by the absorption of one photon in the primary photochemical process. Consequently, photosynthesis correlates better with the number of photons than with energy. McCree209 established that photon (quantum) flux between 400 and 700 nm is a good measure of photosynthesis. Photosynthetically Active Radiation (PAR) is defined in terms of the number of moles of photons between 400 and 700 nm. One mole of photons is approximately 6 × 1023 photons (Avogadro's number). The Photosynthetic Photon Flux Density (PPFD), or the photon irradiance, is expressed in mol/s/m2. Quantum meters with a bandpass from 400 to 700 nm can be used to measure PPFD. For a specific spectral power distribution, PPFD can be related to irradiance or illuminance. For most light sources, the conversion from illuminance to PPFD is in the range between 0.01 to 0.02 µmol/s/m2 per lx.

Figure 5-24. Fluorescent lamps for plant growth have been designed with emission spectra that closely match the absorption spectra of chlorophyll (a in Figure 5-21).

PLANT APPLICATIONS

In plant science there are two main uses for lighting: photosynthesis and photoperiodism. In lighting for photosynthesis, light is applied to plants to sustain, partly or wholly, the photosynthetic processes necessary for desired growth. In lighting for photoperiodism, light is applied to plants to sustain, partly or wholly, the photoperiod necessary to produce a desired flowering response. For many plants, the quantity of light required for photosynthesis can be from 10 to more than 100 times greater than that required for photoperiodism.

Photosynthetic Lighting in Greenhouses Photosynthetic lighting is used in the greenhouse during periods of diminished sunlight in winter months for the growth of out-of-season crops (Figure 5-25). This supplementary lighting can be much less intense than full sunlight, depending on the requirements of the particular plant species. The different applications of lighting for photosynthesis are lighting in the greenhouse before sunrise or after sunset to extend the light period; lighting in the greenhouse on dark, overcast days for the whole light period; and lighting in the middle of the dark period.

Figure 5-25. Commercial lettuce production under a 400-W metal halide lighting canopy. This greenhouse is 100% artificially lighted. This method of plant growth is highly effective in cold climates where days are short. This particular grower is in Newfoundland.

Figure 5-26. Growth chamber for lettuce seedlings that employs water-jacketed high pressure sodium lamps as the sole light source. The first 11 days of lettuce production take place in this germination area. Seedlings develop best under constant lighting conditions. Light intensity is maintained at 50 µmol/s/m2 of PAR during the plants' first 24 hr in the growth chamber. For the remaining 10 days, the light intensity is maintained at 250 µmol/s/m2. The photoperiod (or day length) is 24 hr. Seedlings also require closely controlled specific temperature, relative humidity, carbon dioxide, and irrigation.

Photosynthetic Lighting in Growth Rooms Photosynthetic lighting in these areas includes lighting provided totally by lamps within commercial growth facilities or research chambers. It also includes lighting for plants used for aesthetic purposes in any interior space, including commercial and residential gardening.

Considerable attention has been given to the development of commercial production of salad crops, particularly lettuce, in growth rooms to compete with conventional crop production. Such production of crops is called controlled environment agriculture. Hydroponic (soilless) culture of plants is used in these ventures. Plant growth rooms and chambers are now used extensively for research in agricultural experimental stations, educational institutions, and industrial research laboratories for the growth of plants under controlled environmental conditions (Figure 5-26 to 5-28). The environmental conditions that are controlled and monitored include light quantity and duration, temperature, humidity, and carbon dioxide. A research facility that consists of several plant growth chambers (controlled environment rooms) is called a phytotron or, when combined with rooms for animals, a biotron. The maximum radiant flux that fluorescent systems can produce in conventional applications is 400 µmol/s/m2. To obtain this radiant flux, 1500-mA, 2.4-m (8-ft) T-12 fluorescent lamps can be closely spaced and mounted under a white perforated ceiling through which lamp heat is exhausted. Spaced at uniform intervals between the fluorescent lamps are 60- or 100-W incandescent lamps, which provide the long-wavelength and near-IR component desired for some plants. For photon levels of 1000 µmol/s/m2 or higher, high-intensity discharge (HID) or xenon lamps must be used. Most growth rooms have been fitted with HID lamps because the cost is less and the lamps have much longer lives. Rooms are fitted either with metal halide lamps alone or with a combination of metal halide and high-pressure sodium lamps. The former scheme ensures a more uniform spectral distribution in the room but has lower photosynthetically active radiation, and metal halide lamps degrade faster over time than do high-pressure sodium lamps. Both of these HID lamps, as well as xenon lamps, produce more radiant heat than photosynthetically active radiation (PAR), and serious leaf and soil heating problems result at high photon levels unless special precautions are taken to absorb the radiant heat before it reaches the plants. For closer control, some growth chamber walls are made either of glossy-white, thermally fabricated material, which absorbs lamp heat, or of specular material that reflects more light to the plants. Uniformly spaced groups of lamps are separately circuited to provide several steps of photon levels while maintaining uniform light distribution. To compensate for lumen depreciation, the system should be designed to produce initially higher photon levels than the maximum level. Photosensors can be used to regulate levels as lamps age (see Chapter 27, Lighting Controls). Quantum meters are used to check levels periodically. Most growth chamber fluorescent or HID lighting systems are operated with ballasts located outside the growing area. This arrangement reduces cooling requirements. In many commercial greenhouses, electric HID lighting is used to supplement daylight and extend growing periods (Figure 5-29).

Photoperiodic Lighting In the United States and Canada the greatest use of lighting in plant reproduction is that of photoperiodism to control the out-of-season flowering of certain species of plants that require specific ratios of light to dark periods for flowering. Plants are classified as to the relative length of light period to dark period needed to set flower buds and to bloom. This knowledge is used to bring plants into bloom when there is a particular market need or advantage. In Florida and California the flowering of plants grown for the commercial market is controlled by this type of lighting in the field. During winter months it is essential to extend the day length to promote the flowering of long-day (short-night) plants and to inhibit the flowering of short-day (long-night) plants. It is also essential that the grower be able to shorten the day length to promote the flowering of short-day plants and to inhibit the flowering of long-day plants. During summer months, the grower must apply an opaque cloth or plastic covering over the plants for part of the day to simulate a short day. It is essential that the material used be opaque and that the plants be exposed to no light during this time, because very low levels are effective in this response. The use of lighting and opaque covering permits the growth and flowering of both long-day and short-day plants the year round. Long-day responses for both short-day and long-day plants usually are obtained by irradiating plants 4 to 8 h before sunrise or after sunset or by the more effective 2- to 5-h light period in the middle of the dark period (called a night break).

Figure 5-27. After the lettuce plants have germinated, they are brought to the pond area of the greenhouse where they are given constant light (through conventional 600-W high pressure sodium lamps), irrigation, and nutrients, until they reach full size at approximately 21 days. For the production of 1000 heads of lettuce per day, a 660 m2 (6342 ft2) growing area is required. This greenhouse serves as a "controlled environment agriculture" demonstration site for hydroponic leaf lettuce. Research methods employed here will be applied to commercial greenhouses.

Figure 5-28. The Biomass Production Chamber at Kennedy Space Center, where they are developing systems to raise food in the space colonies of the future. The lighting is high pressure sodium at a very high intensity. The plants in this picture are potatoes.

Figure 5-29. Commercial cucumber production in a northern climate using 1000-W high pressure sodium lamps to supplement daylight. Approximately 1MW of electricity was used per acre. Commercially, the most important group of photoperiodic plants are short-day plants: chrysanthemums and poinsettias. Such plants remain vegetative with a continuous light period greater than 12 h or with a night break. When flowering is

desired, the photoperiod is shortened to approximately 10 h and the night break is discontinued. By providing such longday plants as China aster and Shasta daisy with a continuous 16- to 18-h day as well as supplementary light, they can be brought into flower, whereas continuous short days inhibit flowering. Incandescent and fluorescent lamps are used for photoperiodic lighting. Clear incandescent lamps in industrial reflector luminaires or reflector incandescent lamps are commonly used in the field or in the greenhouse. Some HID and incandescent lamps can produce greater internodal elongation in some plants, and this might be undesirable.

Home and Hobby Applications With available lighting equipment for indoor plant culture, flowering and foliage plants can be taken off the window sill to a place in the room where they can be grown and displayed to best advantage. Some luminaires are equipped with trays to hold moisture to raise the humidity around the plants and with timers to turn lights on and off automatically.

Figure 5-30. Both amateur and professional growers have found value in basement gardens. They allow the amateur to increase the size of the hobby and enable the professional grower to use unproductive space for rooting of cuttings and growth of seedlings. Some amateurs have set up basement gardens of varying sizes in which plants are grown from seed, cuttings, and bulbs as shown in Figure 5-30. A wide variety of flowering and foliage plants, including all plants of the house plant category, are grown under lights. Fluorescent lamps in T-8 and T-12 sizes and with ordinary loadings are acceptable for this type of horticulture. Ordinary loading lamps (400 mA) can be used to grow seedlings and short plants (less than 15 cm [6 in.]) but high-output lamps (800 mA) and very high output lamps (1500 mA) are preferred for taller plants.

Commercial and Industrial Applications Interior spaces with live plants and trees are now commonplace in lobbies, offices, shopping malls, airport waiting rooms, banks, country clubs, restaurants, entryways of condominiums and apartments, and atria of large hotels and of government, commercial, and industrial buildings. Before any plants are chosen or plans are made for their use, the designer, plant specialist, or architect must first determine if the environment is suitable. The photon flux in the space determines the species of plants that survive. If the photon flux due to daylight is not sufficient for live plants, then supplemental lighting should be used, provided that other factors are favorable. The amount of supplemental light is determined by the plant species having the greatest light requirement. Professional plant specialists prefer acclimatized plants for interiorscaping. Acclimatized plants are those that have been conditioned for use in the low-humidity and low-illuminance indoor environments. These plants are taken from greenhouses that often receive full sunlight to a greenhouse with heavy shade. Here they remain for two months or more before being used for interiorscaping. Also, the watering frequency is reduced to condition the plants for indoor use. Such acclimatization prevents the shock that frequently results in rapid defoliation when plants are taken without conditioning from the bright greenhouse to the interiorscape. The following approximate requirements serve as an initial guide in deciding if the environment is suitable for plants.

Specific requirements for each species can differ.

Lighting in interior occupied spaces serves a dual purpose: lighting for people and for plants. Therefore, illuminance is specified rather than irradiance, which is required for controlled growth applications.

COMMON PLANTS AND ILLUMINANCES Approximate recommended illuminances for common plants in interiors are shown in Figure 5-31. Plants are categorized as trees, floor plants, and table or desk plants. The lower illuminances in this figure are the minimum for maintenance. Higher values are more satisfactory for good plant condition. The recommendations are for acclimatized plants receiving 14 h of light per day.

Figure 5-31. Recommended Illuminances* for Acclimatized Plants (14 h of Light Per Day)

Aquarium and Terrarium Lighting225-231 Aquaria and terraria are also found in the home, office, and school for hobby, decorative, and educational purposes. Aquarium lighting serves both a functional and an ornamental purpose when plants are part of the aquarium environment. Through the process of photosynthesis, lighted aquarium plants increase the oxygen level essential for fish respiration and at the same time reduce the carbon dioxide level, preventing the buildup of carbonic acid, which can be harmful to

fish. The light also illuminates both the fish and the aquarium. Dramatic colors of both fish and plants can be observed when fluorescent plant growth lamps (Figure 5-24) are used, because the high long- and short-wavelength output of these lamps emphasizes red and blue colors. Overall, however, colors can appear more natural with conventional rare-earth or daylight fluorescent lamps. Both fluorescent and incandescent lamps are used, with a preference for the former because they are more efficacious. Lighting requirements for aquaria usually range from 0.25 to 0.5 lamp watts per liter of tank capacity when using fluorescent lamps. Lighting can have important behavioral effects on fish in aquaria. Many fish rely heavily on vision for feeding and locating mates.225 The visual sensitivity of fish differs from that of humans and other species. Some fish appear to have visual responses to UV-A radiation that might assist in locating food sources at specific times of the day.226 Light-dark cycling of lighting conditions also can play an important role in breeding routines.227 Terrarium lighting usually requires both fluorescent and incandescent lighting. As a rule of thumb, fluorescent lighting is applied at approximately 200 lamp W/m2 for the plant life while an incandescent lamp is used to light a portion of the terrarium to provide IR radiation for such animals as lizards and frogs, usually found in such environments. Many animals commonly kept in terraria have a parietal eye,228 a light-sensitive region that appears to be involved with circadian regulation. The mating behavior patterns of male reptiles can be disturbed if light-dark cycling is not performed regularly.229 Another important requirement in lighting for terraria containing reptiles or amphibia is to provide sufficient UV radiation to allow for the production of vitamin D.230 This vitamin is essential for skeletal development as well as the visual and immune systems. Caution must be taken, however, to avoid excessive UV radiation, which has been demonstrated to hamper growth and development of eggs in several species.231 The requirements for UV radiation depend on the species, natural habitat, and egg-laying routines.

EFFECTS ON MATERIALS232-240 Fading and Bleaching Fading and bleaching of colored textiles and other materials on exposure to light and other radiant energy is of special interest because of the high illuminances now employed in merchandising. Consequently, a knowledge of some of the factors involved is important. Some of these (not necessarily in order of importance) are as follows:

Illuminance Duration of exposure Spectral distribution of the radiation Moisture Temperature of the material Chemical composition of the dye or other colorant Saturation of the dye (tints versus saturated colors) Composition and weave of fabric Intermittency of exposure Chemical fumes in the atmosphere

While many studies of fading and colorfastness have been published, especially in textile journals, most of them are deficient in data on the illuminances involved. In general, these articles have dealt primarily with improvements in dyes and dyeing methods. Such tests have involved exposures to daylight in various geographical regions and to standardized types of arc lamps, called "fading lamps." The National Institute of Standards and Technology (NIST) has developed standardized methods and lamps for conducting such tests. A long review of research in this field,232 with extensive bibliography, summarizes the subject as follows. The rate at which a dye fades is governed by seven factors:

Photochemistry of the dye molecule Physical state of the dye Chemistry of the substrate Fine structure of the substrate Presence of foreign substances

Atmosphere Irradiance

Certain general conclusions have been derived from studies of several hundred specimens of colored textiles. In view of the fact that the tests have been limited to an infinitesimal percentage of the dyes and textiles in general use, it must be realized that such conclusions are not definitive and many exceptions can be found. The irradiance and duration of exposure are two of the most important factors. Two studies233,234 indicate an approximate reciprocal relationship between time and illuminance in the production of fading; that is, the fading is dependent on the product of these two factors and is substantially unaffected by variations in both as long as the product is unchanged. A third study235 disagrees with this conclusion, indicating that at higher illuminances this relationship breaks down. The spectral power distribution of the incident radiation used affects the rate of fading. It has been found236 that UV-B energy is a small component in most practical light sources but can cause very rapid fading and other forms of product deterioration in some cases. Most electric light sources emit UV-A radiation and this spectral region produces more fading per unit of energy than an equal amount in the visible spectrum. Filters that absorb much of the UV-A but very little visible radiation, have been found to reduce fading somewhat,237 but not as much as is sometimes suggested. It has been shown that fading is produced by energy throughout the entire visible spectrum shorter than approximately 600 nm. Daylight produces more fading than tungsten and fluorescent sources for the same illuminance because daylight has more energy in the short-wavelength region of the spectrum.233 The germicidal (bactericidal) lamp, producing high energy at 253.7 nm, has been used as a potent source for accelerated fading tests236 even though no relationship has been found between fading by germicidal lamps and fading by sunlight or commonly used electric light sources. Figure 5-32 shows spectral reflectance curves for new and slightly faded specimens of pink silk cloth. The spectral changes indicate bleaching in regions of maximum absorption and darkening in regions of minimum absorption. These changes are typical of many specimens tested. Fading appears to be a photochemical process requiring oxygen and is inhibited or greatly reduced in a vacuum. Moisture can also enhance fading. Cellulose is particularly affected by moisture, but wool is less affected, for example.232 Temperature appears to have little effect on the fading rate of silk and cotton at temperatures below 50°C (120°F), but the rate is approximately twice as great at 65°C (150°F) as it is at 30°C (85°F). Often it is found that a light tint is more susceptible than a higher concentration of the same dye. Fading is of major importance to the merchandising field. In one early investigation, tests of approximately 100 textile specimens showed that approximately half of the samples faded to some degree after 500 klx × h (50,000 fc × h) of incandescent illumination. A later study238 of 100 commercial fabrics suggests that several times that exposure is required to produce a minimum perceptible fading with incandescent and fluorescent lamps. Dyes have greatly improved in lightfastness, and fading is not as extensive a problem as it was when the fluorescent lamp was introduced.

Figure 5-32. Spectral reflectance of a specimen of pink silk before and after exposure sufficient to cause moderate fading. Fading of merchandise is most readily apparent where one area has received a high level of exposure and an adjacent area has not. Typical examples are folded neckties and socks stacked on shelves in display cases. To reduce perceptible fading, displayed goods should be rotated periodically. Grocery stores display packaged meats in refrigerated cases under high illuminances. Fresh unprocessed meats show no

appreciable color change due to light within any reasonable display period, although unwrapped meats can show color changes due to dehydration. However, many processed meats, such as bologna, receive their red color from a curing process using salt or sodium nitrate. Through a reaction with light and air, such meats return to their original gray color, and this fading takes place very rapidly in some meats. Some, especially veal loaf and bologna, can show perceptible color change in 1500 to 2000 lx×h (150 to 200 fc×h). Since the illuminance in some display cases may be as high as 1000 lx (100 fc), undesirable changes can occur in 1.5 to 2 h. The most susceptible meats should be placed as far away from the lamps as possible. Depending on the degree of original muscle pigmentation, frozen meat is considered salable for 3 to 6 days under illuminances between 500 and 2000 lx (50 and 200 fc).239 Above 2000 lx (200 fc), shelf life is considerably reduced. Differences in the spectral power distribution of the light sources result in no apparent or statistically significant differences in the rate of color degradation. Cigars can bleach to some degree in cases illuminated by fluorescent lamps, but the exposures must be quite high. In a test using seven brands of cigars it was found that an exposure of approximately 400 klx × h (40,000 fc × h) produced a just-noticeable change and that this exposure could be doubled before the color change reached an objectionable degree. Exposures of this magnitude are greater than typical for the merchandising of cigars.

Luminescence and Luminescent Materials Photoluminescence (see Chapter 1, Light and Optics) occurs in many hundreds of materials when they are exposed to radiation. The most important practical application of photoluminescent materials is in fluorescent light sources, where UV radiation excites the lamp phosphors. These phosphors are oxygen-dominated inorganic crystalline materials. Other materials, such as the zinc and cadmium sulfides and a wide variety of organic compounds excited by UV-A radiation, are used extensively to achieve spectacular theatrical effects and in various signs and instrument dials and low-location emergency lighting. So-called optical bleaches are fluorescent organics used as whiteners in laundered items such as shirts and sheets; they are excited by UV-A and short-wavelength visible radiation to fluoresce, appearing bright blue. This balances for the natural yellow-white appearance of the unimpregnated cloth. Superbright orange and red organic dyes that fluoresce under UV-A excitation are widely used as identification and warning markers, for example, in highspeed aircraft to aid in rapid visual acquisition to avoid collision. Fluorescent paint, ink, and dyed fabrics are available in many colors, including red, orange, yellow, blue, and a white that appears blue under UV radiation. Because these materials transform UV and short-wavelength visible energy into light, as well as reflect incident light, their brightness under daylight is conspicuously high. Some fluorescent materials have an apparent reflectance (under daylight) of over 100%; that is, they emit more visible light than strikes them. These colored fluorescent materials are especially useful on signal flags and signal panels because they can be identified more easily and at greater distances than nonfluorescent surfaces. The increased range over which the fluorescent flags can be identified is most apparent during the half-light conditions of dawn and twilight. Organic fluorescent dyed materials are at times used to produce spectacular signs, such as used on streetcars or buses, and very colorful clothing. Other photoluminescence applications include X-ray- and gamma-ray-stimulable crystals, which find extensive use in scintillation counters, used for detecting and quantifying the exciting radiation. Chemical analyses often are based on the use of the characteristic luminescence of certain activator ions in known host media. Many types of glass fluoresce to some extent. This is important in certain scientific work where even a small fluorescent emission from a glass filter can ruin an experiment. Cathodoluminescent materials find their most important application in television screens and in such scientific instruments as oscilloscopes, electron microscopes, image intensifiers, and radar screens. Here zinc and cadmium sulfides and oxygen-dominated phosphors such as the silicates, phosphates, and tungstates are used. An improvement in color television screens has resulted from the development of a rare-earth (europium) activated deep-red phosphate phosphor. Luminescence attending chemical reactions has been observed widely in both organic and inorganic systems. One of the most interesting is the reaction between the naturally occurring chemicals luciferin, luciferase, and adenosine triphosphate (ATP) within the firefly. Sound and friction of phosphors are phenomena of relatively little practical application. Electric field excitation of phosphors has found use in readout devices and recently in electrodeless fluorescent lamps (see Chapter 6, Light Sources). Phosphorescent Materials. Phosphorescent materials, excited by UV energy, daylight, or light from electric lamps, have been shown to have a high brightness of afterglow for periods of 6 to 9 h, and some of them have a noticeable brightness for as long as 24 h after the exciting source has been removed. Certain phosphorescent materials, generally combinations of zinc, calcium, cadmium, and strontium sulfides, can be incorporated into adhesive tapes (plastic overcoatings), paints, and certain molded plastics. Because of the tendency of many plastics either to transmit moisture, which decomposes the sulfide, or to react directly with the phosphor, care should be exercised in the choice of a plastic to carry the phosphorescent powders. Both vinyl and polystyrene plastics have been found to be well suited to this

application. Phosphorescent materials are suitable only for applications where it is possible to light them before they are needed. Although some materials can be used in spots where a visible brightness is necessary for about 6 to 9 h, only a few of the many phosphorescent compounds have this degree of persistence. Those manufactured from zinc sulfide have high initial brightness after extinguishing the light source, but their useful brightness period does not extend beyond 20 to 30 minutes. Before refinements in the processing of calcium and strontium phosphors were made, the useful brightness of these types did not extend beyond 2 to 3 h after activation. Today, strontium aluminate phosphors can have significantly longer usefulness. Now that long-persistence phosphors are available, phosphorescent materials are suitable for nightlong use in many applications. Brightness reduction (decay) rates are hastened by high temperatures. At very low temperatures (60 K) luminescence may be completely arrested, to recur later upon warming. Radioactive Excitation. This is simply excitation by electrons, ions (atoms, nuclear fragments), or γ-rays, singly or in combination, resulting from the fission or radioactive decay of certain elements. For example, radium emits particles that can excite luminescence when they strike a suitable phosphor. Krypton-85 excites by emission of β-rays or high-energy electrons. The sulfide phosphors emit light not only when exposed to UV energy or light, but also under bombardment by the rays from radioactive materials. Thus, by compounding a mixture of such a radioluminescent material such as zinc sulfide and a small amount of radioactive material, a self-luminous mixture can be produced. Such a radioactive luminous compound continues to emit light without the help of external excitation for a very long time (several years) in practical applications. Radioactive-luminous materials have been used for many years on watch and clock dials and on the faces of other instruments that must be read in the dark. They are the only type of commercially available luminous materials that maintain self-luminosity for years. The power source was formerly some salt of radium or more frequently the lowerpriced mesothorium. These were displaced by strontium-90, and more recently polonium, which has certain advantages with regard to cost and safety. Radioactively powered exit signs typically use tritium, a source of β-rays. See Chapter 29, Emergency, Safety, and Security Lighting, for more information on self-luminous signs. The bombardment of the fluorescent materials by radiation causes them to decompose, which limits the life of the combination. A good-quality material is useful for a few years and appears to maintain a fairly constant brightness during this period. The life of radioactive-luminous paint is controlled to a great extent by the concentration of radioactive material in the mixture, as is the brightness. Increased radioactive content means increased brightness, but more rapid decomposition of the glowing salt. Because of the expense of the radioactive substances used to activate this material, radioactive-luminous paint seldom is used in large quantities or to cover large areas.

EFFECTS OF IR ENERGY241-249 Heat may be transferred from one body to another by conduction, convection or radiation, or by a combination of these processes. IR heating involves energy transfer primarily by radiation, although some convection heating can exist simultaneously due to natural or forced air movement. Transfer of energy or heat occurs whenever radiant energy emitted by one body is absorbed by another. However, only the wavelengths longer than those of visible radiation and shorter than those of radar are used for radiant heating (770 to 100,000 nm). Energy absorption of white, pastel-colored, and translucent materials is best obtained from emissions longer than 2500 nm, whereas the majority of dark-pigmented and oxide-coated materials readily absorb the full range of emissions, visible as well as IR radiation. Water vapor, steam, and other gases absorb IR energy in specific, characteristic bands throughout the spectrum (e.g., Figure 5-1). Glass and quartz materials effectively transmit IR energy only out to about 5000 nm.

Sources of IR Energy Many sources for producing IR energy are available (see Chapter 6, Light Sources). These can be classified generally as point, line, and area sources. Their temperatures, spectral power distribution, and life characteristics vary widely, although source selection generally is not critical unless the products to be heated are selective as to wavelength penetration or absorption, as in the case of many translucent plastics. Tungsten-filament heaters provide essentially instant on-off response from a power source, and their radiant energy efficiency, 86% of power input, makes them a preferred source of IR radiation. Other heaters have thermal inertia varying from approximately 1 min for quartz tubes to 4 to 5 min for metal-sheath heaters. Operating efficiencies are substantially influenced by the design and maintenance of external reflector systems and to a lesser extent by the air temperature and velocity within the heating zone. Overall efficiencies of 35 to 60% are readily obtained in well-designed systems where a long holding time at the designed product temperature is not required. All quartz heat sources can accept high thermal shock. However, metal heaters are best for applications subject to mechanical shock and vibration. A variety of porcelain

holders and terminals is available for these sources. Specular reflectors of anodized aluminum, gold or rhodium are recommended for directing the radiant energy to product surfaces. By comparison, gas IR systems require far heavier and more costly construction to comply with insurance safety requirements. In calculating their operating efficiencies, one must take into account energy loss in the combustion flue products as well as other design factors affecting the ultimate energy utilization.

Product Heating with IR IR radiant energy can be used for any heating application where the principal product surfaces can be arranged for exposure to the heat sources.241 Modern methods of conveying materials have greatly accelerated the use of heat sources arranged in banks or tunnels. Typical applications include:

Drying and baking of paints, varnishes, enamels, adhesives, printer's ink, and other coatings Preheating of thermoplastic materials for forming and tacking Heating of metal parts for shrink-fit assembly, forming, thermal aging, brazing, radiation testing, and conditioning surfaces for application of adhesives and welding Dehydrating of textiles, paper, leather, meat, vegetables, pottery, and sand molds Spot and localized heating for any desired objective

Rapid heating can be provided in relatively cold surroundings by controlling the amount of radiant energy, absorption characteristics of the exposed surfaces, and rate of heat loss to the surroundings.242 Highly reflective enclosures, with or without thermal insulation, are commonly employed to assure maximum energy use. Limited amounts of air movement are often essential in portions or all of the heating cycle, to avoid temperature stratification and assure removal of water or solvent vapors.243 Product temperature control is normally provided by varying the exposure time to IR radiation, or the heater wattage per unit area of facing tunnel area. With modern linear heaters, power densities of 5 to 130 kW/m2 (0.5 to 12 kW/ft2) can accommodate high automation speeds. Where precise temperatures are needed, the design condition may then be modified by voltage or current input controls to add flexibility for a variety of product conditions, handling speeds, or chemical formulations. The temperature of moving parts can be accurately measured by scanning with a radiation pyrometer to provide indication or full automatic control of the heating cycle. Where quality standards permit small variations in temperature, an initial installation test may be made with portable instrumentation, and thereafter the cycle can repeat itself with a degree of reliability consistent with the power supply voltage. This avoids the need for the usual controls required for other types of process heating. Spot heating of a portion of an object can eliminate the need for energy formerly required to preheat the whole object. Appropriate applications of IR heating can lead to more efficient use of energy in production facilities.

Comfort Heating with IR The use of IR radiation for heating in commercial and industrial areas has become quite popular. The T-3 quartz lamp as a semiluminous IR source has distinguished itself for a wide variety of applications in commercial buildings, marquee areas, industrial plants, warehouses, hangars, stadiums, pavilions, and other public areas. Units are usually of the pendant or recessed type, with reflector control for the combined visible and IR radiation. In contrast, residential use (except for bathroom areas) is mostly confined to such low-temperature sources as electric baseboards and plastered radiant ceilings. By supplying heat only when and where needed, IR heating allows thermostats of conventional heating systems to be lowered while comfort is maintained locally and energy conserved overall. Applications. Radiant comfort heating applications fall into two broad classifications: general heating and spot heating.244 General heating installations irradiate complete room areas. High levels of building insulation generally are recommended. The installation in this case often provides a uniform radiant power density in the range of 100 to 320 W/m2 (10 to 30 W/ft2) incident on the floor surface. However, some system designers prefer equipment layouts using asymmetric units to provide a somewhat higher density in the areas adjacent to outside walls to help offset the wall thermal loss. To date, like convection heating systems, overall radiant systems have had an installed capacity sufficient to hold the desired indoor temperature and overcome the building heat loss at the specified outdoor design temperature. However, performance data on some installations indicate that a heating capacity sufficient to overcome only 70 to 90% of the building thermal loss is adequate.245 This reduction probably is due to the direct personnel heating, which elevates the mean radiant temperature in the space. IR radiation passes through air with little absorption, and this is particularly true of the near IR. Therefore, installations involving quartz IR lamps can be mounted at much greater heights than those with far-IR sources. One can choose

equipment with a narrow beam spread so the radiation can be confined primarily to the floor, where it is most beneficial, and losses through the walls are limited. By this means, the mounting height can be increased without requiring a greater installed capacity. It is good practice to keep the radiation from striking the walls at heights more than 2.4 m (8 ft) above the floor. Although IR heating of the air is minimal, air in radiation-heated areas is warmed from energy absorption by the floor and other solid surfaces. Because the heated air rises, room temperature can be controlled with air thermostats245 that are shielded from the IR sources. IR heating systems have an advantage over convection air heating systems for spaces that are subject to high rates of air change (for example, where overhead doors are opened frequently). In these areas, the warm air is lost immediately and air temperature recovery can be lengthy with convection heating. With a radiant system, most objects are warmer than the air, so the air in the space recovers temperature faster. With quartz tubes, metal-sheath heaters, and gas-fired IR units that produce no light, on-off cycling of the equipment is permissible. When lamps that produce light are used, the cycling should be from full to half voltage to prevent large changes in illumination. Certain types of incandescent heat lamps are manufactured with red glass bulbs to minimize light output and the potential for visual distraction. This typically reduces the radiant power by less than 5 percent. Spot Heating. The greatest potential use for high-intensity radiant heating lies in spot or zone heating in exposed areas where conventional heating is impractical, for example, marquees, waiting platforms and loading docks, and such infrequently used areas as stadiums, arenas, viewing stands, houses of worship, and assembly halls. The radiation intensity needed for spot heating varies with a number of factors. The major ones are: 1. The degree of body activity as dictated by the task. The more physical effort expended by the worker, the lower the target temperature. Clothing also influences the target temperature. 2. The lowest temperature that is apt to exist in the space (or the lowest temperature at which the owner wants to provide comfort). 3. The amount of air movement at the location. Indoor drafts and slight air movements outdoors can be overcome by higher irradiances, but compensation for wind velocities of more than 2.2 to 4.5 m/s (5 to 10 mi/h) at temperatures below -1°C (30°F) is not sufficient. Wind screens are far more beneficial than increased radiation levels. For spot heating, units should be positioned to supply radiation from at least two directions, preferably above and from the side of the area to be heated. Care should be taken to avoid locating equipment directly over a person's head. In practice, levels for spot heating vary from 100 W/m2 (10 W/ft2) at waist level for an indoor installation supplementing an inadequate convection system to more than 1 kW/m2 (100 W/ft2) for a marquee or sidewalk people-heating system. At the higher radiation levels, ice and snow are melted246 and water on the floor is evaporated. This can reduce the safety hazard of a slippery floor and improve housekeeping by minimizing the tracking in of snow and water in inclement weather. Where snow melting is desirable, the heating units should be energized as soon as snow starts to fall to avoid any accumulation and consequent high reflection of IR energy. IR heating installations in infrequently used areas can be often turned on before an event to preheat the room surfaces, then turned off before the event is over, with the heat stored in the surfaces and body heat maintaining the comfort level. Many of the control strategies discussed in Chapter 27, Lighting Controls, apply equally to the control of IR heating systems.

REFERENCES 1. Ackerman, B., E. Sherwonit, and J. Williams. 1989. Reduced incidental light exposure: Effect on the development of retinopathy of prematurity in low birth weight infants. Pediatrics 83(6):958-962. 2. Silverman, W. A. 1980. Retrolental fibroplasia: A modern parable. New York: Grune & Stratton. 3. Baerts, W., R. A. Valentin, and P. J. Sauer. 1992. Ophthalmic and cerebral blood flow velocities in preterm infants: Influence of ambient lighting conditions. J. Clin. Ultrasound 20(2):43-48. 4. Boettner, E. A., and J. R. Wolter. 1962. Transmission of the ocular media. Invest. Ophthalmol. 1:176. 5. Berler, D. K. 1989. Muller cell alterations from long-term ambient fluorescent light exposure in monkeys: Ling and

electron microscopic fluorescein and lipofuscin study. Trans. Am. Ophthal. Soc. 87:515-576. 6. Brainard, G., F. M. Barker, R. J. Hoffman, M. H. Stetson, J. P. Hanifin, P. L. Podolin, and M. D. Rollag. 1994. Ultraviolet regulation of neuroendocrine and circadian physiology in rodents. Vision. Res. 34(11):1521-1533. 7. Barker, F. M. and G. C. Brainard. 1991. The direct spectral transmittance of the excised human lens as a function of age, FDA 785345 0090 RA. Washington: Food and Drug Administration. 8. Brainard, G. C., M. D. Rollag, and J. P. Hanifin. 1997. Photic regulation of melatonin in humans: Ocular and neural signal transduction. J. Biolog. Rhythms 12(6):537-546. 9. Bullough J. and M. S. Rea. 1996. Lighting for neonatal intensive care units: Some critical information for design. Light. Res. Tech. 28(4):189-198. 10. Chou, B. R., A. P. Cullen, and K. A. Dumbleton. 1988. Protection factors of ultraviolet-blocking contact lenses. Int. Contact Lens Clin. 15:244-250. 11. Clayman, H. M. 1984. Ultraviolet-absorbing intraocular lenses. Am. Intra-Oc. Imp. Soc. J. 10(4):429-432. 12. Cullen, A. P., K. A. Dumbleton, and B. R. Chou. 1989. Contact lenses and acute exposure to ultraviolet radiation. Opt. Vis. Sci. 66(6):407-411. 13. Dayshaw-Barker, P. 1987. Ocular photosensitization. Photochem. Photobiol. 46(6):1051-1055. 14. Fielder, A. R., J. Robinson, D. E., Y. K. Shaw, Y.K. Ng, and M. J. Mosely. 1992. Light and retinopathy of prematurity: Does retinal location offer a clue? Pediatrics 89(4):648-653. 15. Gies, H. P., C. R. Roy, and G. Elliott. 1990. A proposed UVR protection factor for sunglasses. Clin. Exp. Optom. 73 (6): 184-189. 16. Glass P., G. B. Avery, K.N. Subramanian, M. P. Keyes, A. M. Sostek, and D. S. Friendly. 1985. Effect of bright light in the hospital nursery on the incidence of retinopathy of prematurity. New Eng. J. Med. 313(7):401-404. 17. Goldman, A. I., W. T. Ham Jr., and H. A. Mueller. 1975. Mechanisms of retinal damage resulting from the exposure of rhesus monkeys to ultrashort laser pulses. Exp. Eye Res. 21(5):457-469. 18. Goldman, A. I., W. T. Ham Jr., and H. A. Mueller. 1977. Ocular damage thresholds and mechanisms for ultrashort pulses of both visible and infrared laser radiation in the rhesus monkey. Exp. Eye Res. 24(1):45-56. 19. Goldmann. 1933. Genesis of heat cataract. Arch. Ophthalmol. 9(2):314. 20. Ham, W. T., Jr., R. C. Williams, H. A. Mueller, D. Guerry, A. M. Clarke, and W. J. Geeraets. 1966. Effects of laser radiation on the mammalian eye. Trans. N.Y. Acad. Sci. 28(4): 517-526. 21. Harding, J. J. 1995. The untenability of the sunlight hypothesis of cataractogenesis. Documenta Ophthalmol. 88(3-4): 345-349. 22. Lydahl, E. 1984. Infrared radiation and cataract. Acta Ophthalmol., Suppl. 166. 23. Marshall, J. H. 1991. The susceptible visual apparatus. London: MacMillan. 24. McCanna, P., S. R. Chandra, T. S. Stevens, F. L. Myers, G. de Venecia, and G. H. Bresnick. 1982. Argon laserinduced cataract as a complication of retinal photocoagulation. Arch. Ophthalmol. 100(7):1071-1073. 25. U.S. National Institute for Occupational Safety and Health. 1977. Ocular ultraviolet effects from 295 nm to 335 nm in the rabbit eye. NIOSH 77-1977. Principal investigators D. G. Pitts and A. P. Cullen. Washington DC: National Institute for Occupational Safety and Health. 26. Pitts, D. G., and A. P. Cullen. 1981. Determination of IR radiation levels for acute ocular cataractogenesis. Grafes Arch. Clin. Exp. Ophthalmol. 217(4):285-297.

27. Pitts, D. G., and M. R. Lattimore. 1987. Protection against UVR using the Vistakon UV-Block soft contact lens. Int. Contact Lens Clin. 14:22-29. 28. Pitts, D. G., A. P. Cullen, and P. Dayshaw-Barker. 1980. Determination of ocular threshold levels for infrared radiation cataractogenesis. NIOSH 77-0042-7701. Cincinnati, OH: National Institute for Occupational Safety and Health. 29. Reynolds, J. D., R. J. Hardy, K. A. Kennedy, R. Spencer, W. A. van Heuven, and A. R. Fieldler. 1998. Lack of efficacy of light reduction in preventing retinopathy of prematurity. N. Engl. J. Med. 338(22):1572-1576. 30. Riley, P. A. and T. F. Slater. 1969. Pathogenesis of retrolental fibroplasia. Lancet 2(7614):265. 31. Sanford, B. E., S. Beacham, J. P. Hanifin, P. Hannon, L. Streletz, D. Sliney, and G. C. Brainard. 1996. The effect of ultraviolet-A radiation on visual evoked potentials in the young human eye. Acta Ophthalmol. Scand. 74(6):553-557. 32. Slater, T. F. 1972. Free radical mechanisms in tissue injury. London: Pion. 33. Sliney, D., and M. Wolbarsht. 1980. Safety with lasers and other optical sources. New York: Plenum. 34. Sykes, S. M., W. G. Robinson, M. Waxler, and T. Kuwabara. 1981. Damage to the monkey retina by broad-spectrum fluorescent light. Invest. Ophthalmol. Vis. Sci. 20(4):425-434. 35. Taylor, H. R., S. K. West, F. S. Rosenthal, B. Muñoz, H. S. Newland, H. Abbey, and E. A. Emmett. 1988. Effect of ultraviolet radiation on cataract formation. New Engl. J. Med. 319(22):1429-1433. 36. Waxler, M. 1988. Long-term visual health risks from solar ultraviolet radiation. Ophthalmic Res. 20(3):179-182. 37. Kuwabara, T. 1970. Retinal recovery from exposure to light. Am. J. Ophthalmol. 70(2):187-198. 38. Noell, W. K., and R. Albrecht. 1971. Irreversible effects of visible light on the retina: Role of vitamin A. Science 172: 76-79. 39. Baadsgaard, O. 1991. In vivo ultraviolet irradiation of human skin results in profound perturbation of the immune system. Arch. Dermatol. 127(1):99-109. 40. Bachem, A., and C. I. Reed. 1930. The penetration of ultraviolet light through the human skin. Arch. Phys. Ther. 11 (2): 49-56. 41. Berger, D. 1968. Action spectrum of erythema. In XIII Congressus Internationalis Dermatologiae, edited by W. Jadassohn and C. G. Schirren. Berlin: Springer-Verlag. 42. Coblentz, W. W., and R. Stair. 1934. Data on the spectral erythemic reaction of the untanned human skin to ultraviolet radiation. Bur. Stand. (U.S.) J. Res. 12(1):13-14. 43. Commission Internationale de l'Éclairage. 1987. A reference action spectrum for ultraviolet induced erythema in human skin. CIE Journal 6(1):17-22. 44. Commission Internationale de l'Éclairage. 1993. Reference action spectra for ultraviolet induced erythemal and pigmentation of different human skin types. CIE no. 103/3. Vienna: Bureau Central de la CIE. 45. Cole, C., P. D. Forbes, R. E. Davies, and F. Urbach. 1985. Effect of indoor lighting on normal skin. In The medical and biological effects of light, edited by R. J. Wurtman, M. J. Baum, and J. T. Pott, Annals of the New York Academy of Sciences, vol. 453. New York: New York Academy of Sciences. 46. U.S. Federal Aviation Administration. 1978. On the linkage of solar ultraviolet radiation to skin cancer. FAA EQ-7819. Prepared by P. Cutchis. Springfield, VA: Federal Aviation Administration. 47. Daniels, F., Jr., and B. E. Johnson. 1974. Normal, physiologic and pathologic effects of solar radiation on the skin. In Sunlight and man: Normal and abnormal photobiologic responses, edited by T. B. Fitzpatrick. Tokyo: Univ. of Tokyo Press. 48. Everett, M. A., R. M. Sayre, and R. L. Olson. 1969. Physiologic response of human skin to ultraviolet light. In

Biologic effects of ultraviolet radiation, edited by F. Urbach. Oxford: Pergamon. 49. Freeman, R., D. W. Owens, J. M. Knox, and H. T. Hudson. 1966. Relative energy requirements for an erythemal response of skin to monochromatic wave lengths of ultraviolet present in the solar spectrum. J. Invest. Dermatol. 47(6): 586-592. 50. Kripke, M. L. 1986. Immunology and photocarcinogenesis. J. Am. Acad. Dermatol. 14(1):149-155. 51. Krutmann, J., and C. A. Elmets. 1988. Recent studies on mechanisms in photoimmunology. Photochem. Photobiol. 48(6):787-798. 52. Morrison, W. L. 1989. Effects of ultraviolet radiation on the immune system in humans. Photochem. Photobiol. 50 (5): 515-524. 53. Muel, B., Cersarini J.-P., and J.M. Elwood. 1988. Malignant melanoma and fluorescent lighting. CIE Journal 7 (1):29-32. 54. Pathak, M., and K. Stratton. 1969. Effects of ultraviolet and visible radiation and the production of free radicals in skin. In Biologic effects of ultraviolet radiation, edited by F. Urbach. Oxford: Pergamon. 55. Quevedo, W., Jr. 1974. Light and skin color. In Sunlight and man: Normal and abnormal photobiologic responses, edited by T. B. Fitzpatrick. Tokyo: Univ. of Tokyo Press. 56. Sams, W. M. 1974. Inflammatory mediators in ultraviolet erythema. In Sunlight and man: Normal and abnormal photobiologic responses, edited by T. B. Fitzpatrick. Tokyo: Univ. of Tokyo Press. 57. Holick, M. F. 1989. 1,25-Dihydroxyvitamin D3 and the skin: A unique application for the treatment of psoriasis. Prod. Soc. Exp. Biol. Med. 191(3):246-257. 58. Holick, M. F. 1989. Vitamin D: Biosynthesis, metabolism, and mode of action. Chapter 56 in Endocrinology, vol. 2, edited by L. J. DeGroot. New York: W. B. Saunders. 59. Maclaughlin, J. A., R. R. Anderson, and M. F. Holick. 1982. Spectral character of sunlight modulates photosynthesis of previtamin D3 and its photoisomers in human skin. Science 216:1001-1003. 60. Webb, A. R., L. Kline, and M. F. Holik. 1988. Influence of season and latitude on the cutaneous synthesis of vitamin D3: Exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in human skin. J. Clin. Endocrinol. Metab. 67(2):373-378. 61. Webb, A. R., C. Pilbeam, N. Hanafin, and M. F. Holick. 1990. An evaluation of the relative contributions of exposure to sunlight and of diet to the circulating concentrations of 25-hydroxyvitamin D in an elderly population in Boston. J. Clin. Nutr. 51(6):1075-1081. 62. Akerstedt, T., A. Knuttson, L. Alfredsson, and T. Theorell. 1984. Shiftwork and cardiovascular disease. Scandinavian J. Work and Environmental Health 10:409. 63. Arendt, J. 1995. Melatonin and the mammalian pineal gland. London: Chapman and Hall. 64. Aschoff, J. 1981. A survey on biological rhythms. In Biological rhythms, Handbook of behavioral neurobiology, Volume 4. New York, London: Plenum. 65. Binkley, S. 1990. The clockwork sparrow: Time, clocks, and calendars in biological organisms. Englewood Cliffs, NJ: Prentice-Hall. 66. Brainard, G. C., A. J. Lewy, M. Menaker, R. H. Fredrickson, L. S. Miller, R. G. Weleber, V. Cassone, and D. Hudson. 1988. Dose-response relationship between light irradiance and the suppression of plasma melatonin in human volunteers. Brain Res. 454(1-2):212-218. 67. Rivkees, S., P. Hofman, and J. Fortman. 1997. Newborn primate infants are entrained by low intensity lighting. Proc. Natl. Acad. Sci. USA 94:292-297. 68. Brainard, G. C., B. A. Richardson, T. S. King, and R. J. Reiter. 1984. The influence of different light spectra on the suppression of pineal melatonin content in the Syrian hamster. Brain Res. 294(2):333-339.

69. Bronstein, D. M., G. H. Jacobs, K. A. Haak, J. Neitz, and L. D. Lytle. 1987. Action spectrum of the retinal mechanism mediating nocturnal light-induced suppression of rat pineal gland N-acetyltransferase. Brain Res. 406 (1/2):352-356. 70. Bullough, J., M. S. Rea, and R.G. Stevens. 1996. Light and magnetic fields in a neonatal intensive care unit. Bioelectromagnetics 17(5):396-405. 71. Campbell, S. S., and P. J. Murphy. 1998. Extraocular circadian phototransduction in humans. Science 279:396-399. 72. Cardinali, D. P., F. Larin, and R. J. Wurtman. 1972. Control of the rat pineal gland by light spectra. Proc. Natl. Acad. Sci. (U.S.) 69(8):2003-2005. 73. Czeisler, C. A., T. L. Shanahan, E. B. Klerman, H. Martens, D. J. Brotman, J. S. Emens, T. Klein, and J. F. Rizzo, 3rd. 1995. Suppression of melatonin secretion in some blind patients by exposure to bright light. New. Engl. J. Med. 332 (1): 6-11. 74. Czeisler, C. A., J. S. Allan, S. H. Strogatz, J. M. Ronda, R. Sanchez, C. D. Rios, W. O. Freitag, G. S. Richardson, and R. E. Kronauer. 1986. Bright light resets the human circadian pacemaker independent of the timing of the sleep-wake cycle. Science 233:667-671. 75. Czeisler, C. A., M. C. Moore-Ede, and R. M. Coleman. 1982. Rotating shift work schedules that disrupt sleep are improved by applying circadian principles. Science 217:460-463. 76. Folkard, S., and T. H. Monk. 1985. Hours of work: Temporal factors in work scheduling. New York: Wiley. 77. Halberg, F., E. A. Johnson, B. W. Broun, and J. J. Bittner. 1960. Susceptibility rhythm to E. coli endotoxin and bioassay. Proc. Soc. Exp. Biol. Med. 103(1):142-144. 78. Klein, D. C., R. Y. Moore, and S. M. Reppert, eds. 1991. Suprachiasmatic nucleus: The mind's clock. New York: Oxford University Press. 79. Klein, D. C., R. Smoot, J. L. Weller, S. Higa, S. P. Markey, G. J. Creed, and D. M. Jacobowitz. 1983. Lesions of the paraventricular nucleus area of the hypothalamus disrupt the suprachiasmatic spinal cord circuit in the melatonin rhythm generating system. Brain Res. Bul. 10(5):647-652. 80. Lewy, A. J., R. L. Sack, L. S. Miller, and T. M. Hoban. 1987. Antidepressant and circadian phase-shifting effects of light. Science 235:352-354. 81. Lewy, A. J., T. A. Wehr, F. K. Goodwin, D. A. Newsome, and S. P. Markey. 1980. Light suppresses melatonin secretion in humans. Science 210:1267-1269. 82. Lockley S.W., D.J. Skene, K. Thapan, J. English, D. Ribeiro, I. Haimov, S. Hampton, B. Middleton, M. von Schantz, and J. Arendt. 1998. Extraocular light exposure does not suppress plasma melatonin in humans. J Clin Endocrinol Metab 83(9):3369-3372. 83. Minors, D. S., J. M. Waterhouse, and A. Wirz-Justice. 1991. A human phase-response curve to light. Nuerosci. Lett. 133(1):36-40. 84. Moore, R. Y. 1983. Organization and function of a central nervous system circadian oscillator: The suprachiasmatic hypothalmic nucleus. Federation Proc. 42(11):2783-2789. 85. Moore, R. Y. 1991. The suprachiasmatic nucleus and the circadian timing system. In Suprachiasmatic nucleus, Introduction to Part 2, edited by D. C. Klein, R. Y. Moore, and S. M. Reppert. New York: Oxford Univ. Press. 86. Moore-Ede, M. C., C. A. Czeisler, and G. S. Richardson. 1983. Circadian timekeeping in health and disease. New Engl. J. Med. 309(9):530-536. 87. Moore-Ede, M. C., F. M. Sulzman, and C. A. Fuller. 1982. The clocks that time us: Physiology of the circadian timing system. Cambridge, MA: Harvard Univ. Press. 88. Nelson, D. E., and J. S. Takahashi. 1991. Comparison of visual sensitivity for suppression of pineal melatonin and circadian phase-shifting in the golden hamster. Brain Res. 554(1/2):272-277.

89. Nelson, D. E., and J. S. Takahashi. 1991. Sensitivity and integration in a visual pathway for circadian entrainment in the hamster (Mesocricetus auratus). J. Physiol. 439:115-145. 90. Oren, D. A. 1996. Humoral phototransduction: Blood is a messenger. Neuroscientist 2:207-210. 91. Pickard, G. E., and A. J. Silverman. 1981. Direct retinal projections to the hypothalamus, piriform cortex, and accessory optic nuclei in the golden hamster as demonstrated by a sensitive anterograde horseradish peroxidase technique. J. Comp. Neurol. 196(1):155-172. 92. Podolin, P. C., M. D. Rollag, and G. C. Brainard. 1987. The suppression of nocturnal pineal melatonin in the Syrian hamster: Dose-response curves at 500 and 360 nm. Endocrinology 121(1):266-270. 93. Reiter, R. 1991. Pineal gland: Interface between the photoperiodic environment and the endocrine system. Trends Endocrin. Metab. 2(1):13-19. 94. Ruberg, F. L., D. J. Skene, J. P. Hanifin, M. D. Rollag, J. English, J. Arendt, and G. C. Brainard. 1996. Melatonin regulation in humans with color vision deficiencies. J. Clin. Endocrinol. Metab. 81(8):2980-2985. 95. Stevens, R. G., B. W. Wilson, and L. E. Anderson. 1997. The melatonin hypothesis: Breast cancer and the use of electric power. Columbus, OH: Battelle Press. 96. Takahashi, J. S., P. J. DeCoursey, L. Bauman, and M. Menaker. 1984. Spectral sensitivity of a novel photoreceptive system mediating entrainment of mammalian circadian rhythms. Nature 308:186-188. 97. U.S. Congress. Office of Technology Assessment. 1991. Biological rhythms: Implications for the worker OTA-BA463. Washington: Office of Technology Assessment. 98. Wever, R. A. 1985. Use of light to treat jet lag: Differential effects of normal and bright artificial light on human circadian rhythms. In The medical and biological effects of light, edited by R. J. Wurtman, M. J. Baum, and J. T. Potts, Annals of the New York Academy of Science, 453. New York: New York Academy of Sciences. 99. Wetterberg, L., ed. 1993. Light and biological rhythms in man. New York: Pergamon Press. 100. American Academy of Peridatrics. 1994. Practice parameter: Management of hyperbilirubinemia in the healthy term newborn. Pediatrics 94(4):558-565. 101. Gartner, L. M., C. T. Herrarias, and R. H. Sebring. 1998. Practice patterns in neonatal hyperbilirubinemia. Pediatrics 101(1):25-31. 102. Epstein, J. H. 1989. Photomedicine. In The science of photobiology, edited by K. C. Smith. New York: Plenum. 103. 1995. Fiberoptic phototherapy systems: Clinical and technical overview of phototherapy. Health Devices 24 (12):496498. 104. Assembly of Life Sciences. Committee on Phototherapy in the Newborn. 1974. Phototherapy in the newborn: An overview, edited by G. B. Odell, R. Schaffer, and A. P. Simopoulos. Washington: National Academy of Sciences. 105. Regan, J. D., and J. A. Parrish, eds. 1978. The science of photomedicine. New York: Plenum. 106. Sisson, T. R. C. 1976. Visible light therapy of neonatal hyperbilirubinemia. Chapter 6 in Photochemical and photobiological reviews, vol. 1, edited by K. C. Smith. New York: Plenum. 107. Adrian, R. M., J. A. Parrish, T. K. Momtaz, and M. J. Karlin. 1981. Outpatient phototherapy for psoriasis. Arch. Dermatol. 117(10):623-626. 108. Anderson, T. F., T. P. Waldinger, and J. J. Voorhees. 1984. UV-B phototherapy. Arch. Dermatol. 120(11):15021507. 109. Parrish, J. A., and K. F. Jaenicke. 1981. Action spectrum for phototherapy of psoriasis. J. Invest. Dermatol. 76 (5):359362.

110. Bruynzeel, I., W. Bergman, H. M. Hartevelt, C. C. Kenter, E. A. Van de Velde, A. A. Schothorst, and D. Suurmond. 1991. "High single-dose" European PUVA regimen also causes an excess of non-melanoma skin cancer. Brit. J. Dermatol. 124(1):49-55. 111. Gilchrest, B. A., J. A. Parrish, L. Tanenbaum, H. A. Haynes, and T. B. Fitzpatrick. 1976. Oral methoxsalen photochemotherapy of mycosis fungoides. Cancer 38(2): 683-689. 112. Lerman, S., M. Jocoy, and R. F. Borkman. 1977. Photosensitization of the lens by 8-methoxypsoralen. Invest. Ophthalmol. Vis. Sci. 16(11):1065-1068. 113. Lerman, S., and R. F. Borkman. 1977. A method for detecting 8-methoxypsoralen in the ocular lens. Science 197:12871288. 114. Morison, W. L., J. A. Parrish, and T. B. Fitzpatrick. 1978. Oral psoralen photochemotherapy of atopic eczema. Brit. J. Dermatol. 98(1):25-30. 115. Parrish, J. A., T. B. Fitzpatrick, M. A. Pathak, and C. Shea. 1976. Photochemotherapy of vitiligo: Oral psoralen and a new high-intensity long-wave ultraviolet light system. Arch. Dermatol. 112(11):1531-1534. 116. Parrish, J. A., T. B. Fitzpatrick, L. Tanenbaum, and M. A. Pathak. 1974. Photochemotherapy of psoriasis with oral methoxsalen and longwave ultraviolet light. New Engl. J. Med. 291(23):1207-1211. 117. Parrish, J. A., M. J. LeVine, W. L. Morison, E. Gonzalez, and T. B. Fitzpatrick. 1979. Comparison of PUVA and beta-carotene in the treatment of polymorphous light eruption. Brit. J. Dermatol. 100(2):187-191. 118. Pathak, M. A., D. M. Kramer, and T. B. Fitzpatrick. 1974. Photobiology and photochemistry of furocoumarins (psoralens). In Sunlight and man: Normal and abnormal photobiologic responses, edited by T. B. Fitzpatrick. Tokyo: Univ. of Tokyo Press. 119. IES. Photobiology Committee. 1979. Risks associated with use of UV-A irradiators being used in treating psoriasis and other conditions. Light. Des. Appl. 9(3):56-60. 120. Stern, R. S., and R. Lange. 1988. Members of the Photochemotherapy Follow-up Study. Non-melanoma skin cancer occurring in patients treated with PUVA five to ten years after first treatment. J. Invest. Dermatol. 91(2):120-124. 121. Stern, R. S. 1990. Members of the photochemotherapy follow-up study. Genital tumors among men with psoriasis exposed to psoralens and ultraviolet A radiation (PUVA) and ultraviolet B radiation. New Engl. J. Med. 322(16):10931097. 122. Stern, R. S., L. A. Thibodeau, R. A. Kleinerman, J. A. Parrish, T. B. Fitzpatrick, and 22 participating investigators. 1979. Risk of cutaneous carcinoma in patients treated with oral methoxsalen photochemotherapy for psoriasis. New Engl. J. Med. 300(15):809-813. 123. Doiron, D. R., and G. S. Keller. 1986. Porphyrin photodynamic therapy: Principles and clinical applications. In Therapeutic photomedicine, edited by H. Hönigsmann, and G. Stengl. Basel: Karger. 124. Doiron, D. R., L. O. Svaasand, and A. E. Profio. 1983. Light dosimetry in tissue: Application to photoradiation therapy. In Porphyrin photosensitization, edited by D. Kessel, and T. J. Dougherty, Advances in Experimental Medicine and Biology, vol. 160. New York: Plenum. 125. Dougherty, T. J., W. R. Potter, and K. R. Weishaupt. 1984. The structure of the active component of hematoporphyrin derivative. In Porphyrin Localization and Treatment of Tumors, edited by D. R. Doiron and C. J. Gomer, Progress in Clinical and Biological Research, vol. 170. New York: Alan R. Liss. 126. Epstein, J. H. 1989. Photomedicine. Chapter 6 in The science of photobiology, edited by K. C. Smith. New York: Plenum. 127. Regan, J. D., and J. A. Parrish, eds. 1978. The science of photomedicine. New York: Plenum. 128. American Psychiatric Association. 1998. Diagnostic and statistical manual of mental disorders. 4 ed. Washington: American Psychiatric Association.

129. Avery, D., M. A. Bolte, S. R. Dager, L. G. Wilson, M. Weyer, G. B. Cox, and D. L. Dunner. 1993. Dawn simulation treatment of winter depression: A controlled study. Am. J. Psychiatry 150(1):113-117. 130. Brainard, G. C., D. Sherry, R. G. Skwerer, M. Waxler, K. Kelly, and N. E. Rosenthal. 1990. Effects of different wavelengths in seasonal affective disorder. J. Affect. Disord. 20(4):209-216. 131. Lam, R. W., ed. 1998. Beyond seasonal affective disorder: Light treatment for SAD and non-SAD disorders. Washington: American Psychiatric Press. 132. Lewy, A. J., R. L. Sack, L. S. Miller, and T. M. Hoban. 1987. Antidepressant and circadian phase-shifting effects of light. Science 235:352-354. 133. Lewy, A. J., H. A. Kern, N. E. Rosenthal, and T. A. Wehr. 1982. Bright artificial light treatment of a manicdepressive patient with a seasonal mood cycle. Am. J. Psychiatry 139(11):1496-1498. 134. Oren, D. A., G. C. Brainard, S. H. Johnston, J. R. Joseph-Vanderpool, E Sorek, and N. E. Rosenthal. 1991. Treatment of seasonal affective disorder with green light versus red light. Am. J. Psychiatry 148(4):509-511. 135. Rosen, L. N., S. D. Targum, M. Terman, M. J. Bryant, H. Hoffman, S. F. Kasper, J. R. Hamovit, J. P. Docerty, B. Welch, and N. E. Rosenthal. 1990. Prevalence of seasonal affective disorder at four latitudes. Psychiatry Res. 31(2): 131-144. 136. Rosenthal, N. E., D. A. Sack, J. C. Gillin, A. J. Lewy, F. K. Goodwin, Y. Davenport, P. S. Mueller, D. A. Newsome, and T. A. Wehr. 1984. Seasonal affective disorder: A description of the syndrome and preliminary findings with light therapy. Arch. Gen. Psychiatry 41(1):72-80. 137. Rosenthal, N. E., D. A. Sack, R. G. Skwerer, F. M. Jacobsen, and T. A. Wehr. 1988. Phototherapy for Seasonal Affective Disorder. J. Biol. Rhythms 3(2):101-120. 138. Rosenthal, N. E., D. E. Moul, C. J. Hellekson, D. A. Oren, A. Frank, G. C. Brainard, M. G. Murray, and T. A. Wehr. 1993. A multicenter study of the light visor for seasonal affective disorder: No difference in efficacy found between two different intensities. Neuropsychopharmacology 8(2):151. 139. Rosenthal, N. E. 1993. Diagnosis and treatment of seasonal affective disorder. JAMA 270(22):2717-2720. 140. Society for Light Treatment and Biological Rhythms. 1991. 1991 membership directory. Wilsonville, OR: Society for Light Treatment and Biological Rhythms. 141. Stewart, K. T., J. R. Gaddy, B. Byrne, S. Miller, and G. C. Brainard. 1991. Effects of green or white light for treatment of seasonal depression. Psychiatry Res. 38(3):261-270. 142. Stewart, K. T., J. R. Gaddy, D. M. Benson, B. Byrne, K. Doghramji, and G. C. Brainard. 1990. Treatment of winter depression with a portable, head-mounted phototherapy device. Prog. Neuropsychopharmacol. Biol. Psychiatry 14(4): 569-578. 143. Terman, J. S., M. Terman, D. Schlager, B. Rafferty, M. Rosofsky, M. J. Link, P. F. Gallin, and F. M. Quitkin. 1990. Efficacy of brief intense light exposure for treatment of winter depression. Psychopharmacol. Bul. 26(1):3-11. 144. Terman, M., J. S. Terman, F. M. Quitkin, and P. J. McGrath. 1989. Light therapy for seasonal affective disorder: A review of efficacy. Neuropsychopharmacology 2(1):1-22. 145. Terman, M., D. Schlager, S. Fairhurst, and B. Perlman. 1989. Dawn and dusk simulation as a therapeutic intervention. Biol. Psychiatry 25(7):966-970. 146. Wehr, T. A., R. G. Skwerer, F. M. Jacobsen, D. A. Sack, and N. E. Rosenthal. 1987. Eye versus skin phototherapy of seasonal affective disorder. Am. J. Psychiatry 144(6):753-757. 147. Yerevanian, B. I., J. L. Anderson, L. J. Grota, and M. Bray. 1986. Effects of bright incandescent light on seasonal and nonseasonal major depressive disorder. Psychiatry Res. 18(4):355-364. 148. Badia, P., B. Myers, M. Boecker, J. Culpepper, and J. R. Harsh. 1991. Bright light effects on body temperature, alertness, EEG and behavior. Physiol. Behav. 50(3):583-588. 149. Boyce, P. R., J. W. Beckstead, N. H. Eklund, R. W. Strobel, and M. S. Rea. 1997. Lighting the graveyard shift: The

influence of a daylight-simulating skylight on the task performance and mood of night shift workers. Light. Res. Technol. 29(3):105-134. 150. Brainard, G. C., J. P. Hanifin, P. R. Hannon, W. Gibson, J. French, and M. D. Rollag, 1996. The biological and behavioral effects of light in humans: From basic physiology to application. In Biologic effects of light, edited by M. F. Hollick and E. G. Jung. New York: Walter de Gruyter. 151. Dawson, D., and S.S. Campbell. 1991. Timed exposure to bright light improves sleep and alertness during simulated night shifts. Sleep. 14(6):511-516. 152. Czeisler, C. A., M. P. Johnson, J. F. Duffy, E. N. Brown, J. M. Ronda, and R. E. Kronauer. 1990. Exposure to bright light and darkness to treat physiologic maladaptation to night work. New Engl. J. Med. 322(18):1253-1259. 153. Daan, S., and A. J. Lewy. 1984. Scheduled exposure to daylight: A potential strategy to reduce "jet lag" following transmeridian flight. Psychopharmacol. Bul. 20(3):566-568. 154. Dollins, A. B., H. J. Lynch, R. J. Wurtman, M. H. Deng, and H. R. Lieberman. 1993. Effects of illumination on human nocturnal serum melatonin levels and performance. Physiol. Behav. 53(1):153-160. 155. Eastman, C. I. 1990. Circadian rhythms and bright light: Recommendations for shift work. Work and Stress 4 (3):245-260. 156. Eastman, C. I. 1991. Squashing versus nudging circadian rhythms with artificial bright light: Solutions for shift work? Perspect. Biol. Med. 34(2):181-195. 157. French, J., P. R. Hannon, and G. C. Brainard. 1990. Effects of bright illuminance on body temperature and human performance. In Ann. Rev. Chronopharmacol., Vol. 7. Oxford: Pergamon. 158. Monk, T. H., M. L. Moline, and R. C. Graeber. 1988. Inducing jet lag in the laboratory: Patterns of adjustment to an active shift routine. Aviat. Space Environ. Med. 59(8):703-710. 159. Society for Light Treatment and Biological Rhythms. 1991. Consensus statements on the safety and effectiveness of light therapy of depression and disorders of biological rhythms. Light Treat. Biol. Rhythms 3:4-9. 160. Wever, R. A. 1985. Use of light to treat jet lag: Differential effects of normal and bright artificial light on human circadian rhythms. In The medical and biological effects of light, edited by R. J. Wurtman, M. J. Baum, and J. T. Potts, Annals of the New York Academy of Science, 453. New York: New York Academy of Sciences. 161. Eastman, C. I. 1990. What the placebo literature can tell us about light therapy for SAD. Psychopharmacol. Bul. 26 (4):495-504. 162. Ross, M., and J. M. Olson. 1981. An expectancy-attribution model of the effects of placebos. Psychol. Rev. 88(5): 408-437. 163. Karn, T. 1988. Molecular mechanisms of therapeutic effects of low-intensity laser radiation. Lasers Life Sci. 2:5374. 164. Smith, K. C. 1991. The photobiological basis of low level laser radiation therapy. Laser Therapy 3:19-24. 165. Stephenson, C. G., D. S. Gartry, D. O'Brart, M. G. Kerr-Muir and J. Marshall. 1998. Photorefractive keratectomy: A 6-year follow-up study. Ophthalmol. 105(2):273-281. 166. Bickford, E. D., G. W. Clark, and G. R. Spears. 1974. Measurement of ultraviolet irradiance from illuminants in terms of proposed public health standards. J. Illum. Eng. Soc. 4(1):43-48. 167. American National Standards Institute. 1976. American national standard for the safe use of lasers, ANSI Z136.11976. New York: American National Standards Institute. 168. Sliney, D., and M. Wolbarsht. 1980. Safety with lasers and other optical sources. New York: Plenum. 169. U. S. Army Environmental Hygiene Agency. 1979. Laser hazards bibliography. Prepared by D. H. Sliney, N. Krial, D. W. Griffis, and L. L. Ryan. Aberdeen Proving Ground, MD: Army Environmental Hygiene Agency.

170. Sliney, D. H. 1972. The merits of an envelope action spectrum for ultraviolet radiation exposure criteria. Am. Industr. Hyg. Assoc. J. 33(10):644-653. 171. American Conference of Governmental Industrial Hygienists. 1979. Threshold limit values and for physical agents. Cincinnati, OH: American Conference of Governmental Industrial Hygienists. 172. Bergman, R. S., T. G. Parham, and T. K. McGowan. 1995. UV emmission from general lighting lamps. J. Illum. Eng. Soc. 24(1):13-24. 173. ANSI/IESNA RP-27.1-96. 1997. Recommended Practice for Photobiological Safety for Lamps and Lamp Systems: General Requirements. 174. ANSI/IESNA RP-27.3-96. 1997. Recommended Practice for Photobiological Safety for Lamps: Risk Group Classification and Labeling 175. Levin, R. E. 1998. Photobiological safety and risk: ANSI/ IESNA RP-27 series. J. Illum. Eng. Soc. 27(1): 136-143. 176. Brickner, P. W., J. M. McAdam, and L. L. Scharer. 1993. Tuberculosis in homeless populations. In Tuberculosis: A Comprehensive International Approach, edited by L. B. Reichman and E. S. Herschfield. New York: Marcel Dekker, Inc. 177. DeFabo, E. C., and Noonan, F. P. 1983. Mechanism of immune suppression by ultraviolet irradiation in vivo: Evidence for the existence of a unique photoreceptor in skin and its role in immunology. J. Ex. Medicine 158:84-98. 178. Electric Power Research Institute. 1995. TechCommentary: Engineering Controls for Infectious Airborne Organisms, EPRI publication TC106000. Palo Alto, CA: EPRI. 179. Electric Power Research Institute. 1997. Tuberculosis: Infection Controls Help Protect Hospital Staff and Patients, EPRI TB-107739. Palo Alto, CA: EPRI. 180. Friedberg, E. C., K. H. Cook, J. Duncan, and K. Mortelmans. 1977. DNA repair enzymes in mammalian cells. Chapter 5 in Photochemical and photobiological reviews, vol. 2, edited by K. C. Smith. New York: Plenum. 181. Harm, W., C. S. Rupert, and H. Harm. 1971. The study of photoenzymatic repair of UV lesions in DNA by flash photolysis. Chapter 7 in Photophysiology: Current topics in photobiology and photochemistry, vol. 6, edited by A. C. Giese. New York: Academic Press. 182. Kelner, A. 1949. Effect of visible light on the recovery of Streptomyces Griseus Conidia from ultra-violet irradiation injury. Proc. Natl. Acad. Sci. (U.S.) 35(2):73-79. 183. GTE. Sylvania. Germicidal and short wave radiation. Sylvania Engineering Bulletin 0-342. Prepared by C. C. Mpelkas. Danvers, MA: Sylvania. 184. Riley, R. L., and E. A. Nardell. 1989. Clearing the air, the theory and application of ultraviolet air disinfection. Am. Rev. Respir. Dis. 139 (5):1286-1294. 185. Riley, R. L., and E. A. Nardell. 1993. Controlling transmission of tuberculosis in health care facilities: Ventilation, filtration and ultraviolet air disenfection. Plant, Technology and Safety Series: Controlling Occupational Exposure to Tuberculosis. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations. 186. Riley, R. L. 1994. Ultraviolet air disinfection: Rationale for whole building irradiation. Infect. Control Hosp. Epidemiol. 15(5):324-328. 187. Noonan, F., and DeFabo, E. 1992. Immunosuppression by ultraviolet B radiation: Initiation by urocanic acid. Immun. Today 13(7):250-254. 188. Setlow, J. K. 1966. The molecular bases of biological effects of ultraviolet radiation and photoreactivation. Chapter 4 in Current topics in radiation research, vol. 2, edited by M. Ebert, and A. Howard. Amsterdam: North-Holland. 189. Smith, K. C. 1978. Multiple pathways of DNA repair in bacteria and their roles in mutagenesis. Photochem. Photobiol. 28(2):121-129. 190. Sliney, D. H. 1990. Ultraviolet radiation and the eye. In Light, lasers and synchrotron radiation, edited by M. Grandolfo. New York: Plenum Press.

191. Snapka, R. M., and C. O. Fuselier. 1977. Photoreactivating enzyme from Escherichia coli. Photochem. Photobiol. 25 (5): 415-420. 192. Baker, H., and T. E. Hienton. 1952. Traps have some value. In Insects: Yearbook of agriculture 1952. U.S. Department of Agriculture. Washington: U.S. G.P.O. 193. Barrett, J. R., Jr., R. T. Huber, and F. W. Harwood. 1973. Selection of lamps for minimal insect attraction. Trans. Am. Soc. Ag. Eng. 17(4):710-711. 194. Barrett, J. R., Jr., R. A. Killough, and J. G. Hartsock. 1974. Reducing insect problems in lighted areas. Trans. Am. Soc. Ag. Eng. 17(2):329-330, 338. 195. Goldsmith, T. H. 1961. The color vision of insects. In A symposium on light and life, W. D. McElroy and B. Glass. Baltimore, MD: John Hopkins Univ. Press. 196. Hollingsworth, J. P., and A. W. J. Hartstack. 1971. Recent research on light trap design. American Society of Agriculture Engineering Paper, 71-803. 197. U.S. Department of Agriculture. Agricultural Research Service. 1963. Electric insect traps for survey purposes. ARS 42-3-1. Prepared by J. P. Hollingsworth, J. G. Hartsock, and J. M. Stanley. Washington: Agricultural Research Service. 198. U.S. Department of Agriculture. Agricultural Research Service. 1961. Response of insects to induced light: Presentation papers. ARS 20-10. Washington: Agricultural Research Service. 199. Canadian Standards Association. 1998. Canadian electrical code:Part I safety standard for electrical installations, CSA C22.1-1998. Rexdale ON: Canadian Standards Association. 200. North, M.O. 1990. Commercial Chicken Production Manual, 4th ed. New York: Chapman & Hall. 201. Bickford, E. D., and S. Dunn. 1972. Lighting For Plant Growth. 1st ed. Kent, OH: Kent State University Press. 202. Bickford, E. D. 1977. Interiorscape lighting. Light Des. Appl. 7(10):22-25. 203. Langhans, R. W., ed. 1978. Growth chamber manual: Enviromental control for plants. Ithaca, NY: Comstock. 204. Cathey, H. M. 1969. Guidelines for the germination of annual, pot plant and ornamental herb seeds. Part 3. Florists' Rev. 144(3744):26, 29, 75-77. 205. Williams, T. J., W. J. Doty, and A. C. Sinnes. 1982. Gardening under glass and lights. Mount Vernon, VA: American Horticultural Society. 206. Hart, J. W. 1988. Light and plant growth. London: Unwin Hyman. 207. Cathey, H. M., and L. E. Campbell. 1974. Lamps and lighting: A horticultural view. Light Des. Appl. 4(11):41-52. 208. Downs, R. J. 1975. Controlled environments for plant research. New York: Columbia University Press. 209. McCree, K. J. 1972. Test of current definitions of photosynthetically active radiation against leaf photosynthesis data. Agric. Meterol. 10(6):443-453. 210. Cathey, H. M., L. E. Campbell, and R. W. Thimijan. 1978. Plant growth under fluorescent lamps: Comparative development of 11 species. Florists' Rev. 163(4213): 26-29, 67-69. 211. Elbert, G., and Elbert, V. F. 1974. Plants that really bloom indoors. New York: Simon and Schuster. 212. Fitch, C. M. 1972. The complete book of house plants. New York, Hawthorn. 213. Gaines, R. L. 1977. Interior plantscaping: Building design for interior foliage plants, 1st ed. New York: Architectural Record Books. 214. Associated Landscape Contractors of America. Interior Plantscape Division. 1988. Guide to interior landscape

specifications. Falls Church, VA: Associated Landscape Contractors of America. 215. Kranz, F. H., and J. L. Kranz. 1971. Gardening indoors under lights, New rev. ed. New York: Viking. 216. Whately, J. M., and F. R. Whately. 1980. Light and plant life. London: E. Arnold. 217. Orans, M. 1984. Houseplants and indoor landscaping. Clearwater, FL: A. B. Morse. 218. Scrivens, S. and L. Pemberton. 1980. Interior planting in large buildings. New York: Wiley. 219. Withrow, R. B., ed. 1959. Photoperiodism and related phenomena in plants and animals, Publication 55. Washington: American Association for the Advancement of Science. 220. Rabinowitch, E., and Govindjee. 1969. Photosynthesis. New York: Wiley. 221. Austin, R. L. 1985. Designing the interior landscape. New York: Van Nostrand Reinhold. 222. Shibles, R. 1976. Terminology pertaining to photosynthesis: Report by the Crop Science Committee on Crop Terminology. Crop Sci. 16(3):437-439. 223. Sager, J. C., O. W. Smith, J. L. Edwards, and K. L. Cyr. 1988. Photosynthetic efficiency and phytochrome photoequilibria determination using spectral data. Trans. ASAE 31(6):1882-1889. 224. McCree, K. J. 1972. The action spectrum, absorptance and quantum yield of photosynthesis in crop plants. Agric. Meteorol. 9(3/4):191-216. 225. Reimchen, T. E. 1989. Loss of nuptial color in threespine sticklebacks (Gasterosteus aculeatus). Evolution 43 (2):450460. 226. Hawryshyn, C. W., and R. Beauchamp. 1985. Ultraviolet photosensitivity in goldfish: An independent U.V. retinal mechanism. Vision Research 25(1):11-20. 227. McFarland, W. N. 1986. Light in the sea--Correlations with behaviors of fishes and invertebrates. American Zoologist 26(2):389-401. 228. Dodt, E. 1973. The parietal eye (pineal and parietal organs) of lower vertebrates. Chapter 16 in Central processing of visual information B: Visual centers of the brain, VII/3, edited by R. Jung, Handbook of sensory physiology: Berlin: Springer-Verlag. 229. Bartholomew, G. A. 1959. Photoperiodism in reptiles. In Photoperiodism and related phenomena in plants and animals, edited by R. B. Withrow. Washington: American Association for the Advancement of Science. 230. Bidmon, H.-J., and W. E. Stumpf. 1995. 1,25-Dihydroxyvitamin D3 binding sites in the eye and associated tissues of the green lizard (Anolis carolinensis). Histochemical Journal 27(7):516-523. 231. Blaustein, A. R. 1994. Amphibians in a bad light. Natural History 103(10):32-37. 232. Giles, C. A., and R. B. McKay. 1963. The light-fastness of dyes, a review. Textile Res. J. 33(7):527-577. 233. Luckiesh, M., and A. H. Taylor. 1940. Fading of dyed textiles by radiant energy. Am. Dyest. Rep. 29(21):543-546, 548. 234. Luckiesh, M., and A. H. Taylor. 1925. Fading of colored materials by daylight and artificial light. Trans. Illum. Eng. Soc. 20(10):1078-1099. 235. American Association of Textile Chemists and Colorists. Committee on Color-Fastness to Light. 1957. A study of the variables in natural light fading. Am. Dyest. Rep. 46(23): 861-883. 236. DeLaney, W. B., and A. Makulec. 1963. A review of the fading effects of modern light sources on modern fabrics. Illum. Eng. 58(11):676-684.

237. Taylor, A. H. 1946. Fading of colored textiles. Illum. Eng. 41(1):35-38. 238. Taylor, A. H., and W. G. Pracejus. 1950. Fading of colored materials by light and radiant energy. Illum. Eng. 45(3): 149-151. 239. Hansen, L. J., and H. E. Sereika. 1969. Factors affecting color stability of prepackaged frozen fresh beef in display cases. Illum. Eng. 64(10):620-624. 240. Little, A. H. 1964. The effect of light on textiles. J. Soc. Dyers Col. 80(10):527-534. 241. Hall, J. D. 1947. Industrial applications of infrared, 1st ed. New York: McGraw-Hill. 242. Garber, H. J., and F. M. Tiller. 1950. Infrared radiant heating. Ind. Eng. Chem. 42(3):456-463. 243. National Fire Protection Association. 1969. Standards for Class A ovens and furnaces (including industrial infrared heating systems). NFPA 86A. Boston: National Fire Protection Association. 244. Frier, J. P., and W. R. Stephens. 1962. Design fundamentals for space heating with infrared lamps. Illum. Eng. 42 (12): 779-784. 245. 1962. Heating with infrared. Elec. Constr. Maint. 61(8): 92-95, 61(10):133-135. 246. Frier, J. P. 1964. Design requirements for infrared snow melting systems. Illum. Eng. 59(10):686-693. 247. Goodell, P. H. 1941. Radiant heat: A full-fledged industrial tool. Trans. Am. Inst. Elec. Eng. 60:464-470. 248. Bennett, H. J., and H. Haynes. 1940. Paint baking with near infra-red. Chem. Metal. Eng. 47(2):106-108. 249. Haynes, H. 1941. The use of radiant energy for the application of heat. Illum. Eng. 36(1):61-78.

(Video) SLL Lighting Handbook

6 Light Sources A BRIEF HISTORY OF LIGHT SOURCES1,2,3 The earliest man-made light sources were fire, torches, and candles. Ancient Egyptians used hollowed-out stones filled with fat, with plant fibers as wicks. These were the first candles, and they date back to about 3000 BC. In the Middle Ages, candles were made of tallow, a type of animal fat; later they were made of beeswax or paraffin. Modern candles can still be thought of as a type of fat lamp, but their use today is almost entirely decorative. Ancient Greeks and Romans made lamps from bronze or pottery that burned olive oil or other vegetable oils in their spouts. Many oil lamps appeared during the Middle Ages, when reflectors were added to their designs. Early American colonists used fish oil and whale oil in their Betty lamps. Many improvements were made in the design and fabrication of these lamps over the years, but none produced light efficiently until 1784, when a Swiss chemist named Argand invented a lamp that used a hollow wick to allow air to reach the flame, resulting in a bright light. Later, a glass cylinder was added to the Argand lamp allowing the flame to burn better. With the birth of the petroleum industry, kerosene became a widely used fuel in these lamps. In the 1800s gas lamps became popular as street lights, originating in London, England. The gas lamp had no wick, but its chief drawback was an open flame that produced considerable flicker. The electric lamp replaced gas lamps in the late 1800s and early 1900s. The first electric lamp was the carbon-arc lamp, demonstrated in 1801 by Sir Humphrey Davy, but electric lights became popular only after the incandescent lamp was developed independently by Sir Joseph Swan in England and Thomas Edison in the United States. The latter patented his invention in 1879 and subsequently made the invention the commercial success that it is today. Figure 6-1 illustrates the history of different light sources. This century has seen a huge increase in the number of available light sources in the marketplace, starting with improvements in the Edison lamp, then the introduction of mercury vapor lamps in the 1930s, followed closely by fluorescent lamps at the 1939 World Fair. Tungsten-halogen lamps were introduced in the 1950s; metal halide and high pressure sodium (HPS) lamps in the 1960s. The introduction of electrodeless lamps in the 1990s is an indication that the industry is dynamic, and the introduction of new light sources is expected to continue at least at the present rate well into the next century. Figure 6-2 lists the approximate luminance of various light sources. This chapter describes the various light sources and control equipment now available. Fundamental information concerning the generation of light and the operating principles of electric light sources is given in Chapter 1, Light and Optics, and in other references.1 The sun and sky as light sources are covered in Chapter 8, Daylighting. For techniques on the measurement of light and light output from light sources, see Chapter 2, Measurement of Light and Other Radiant Energy. Some specialty lamps are discussed in Chapter 15, Theater, Television, and Photographic Lighting, and data on UV radiation from some common sources are found in the museum section in Chapter 14, Lighting for Public Places and Institutions. With such a wide selection of light sources on the market, it is likely that several different choices could be made for a given lighting application. While the general characteristics can be provided, a definitive list with absolute values for all types and manufacturers would be too extensive for this chapter. Figure 6-3 provides a comparison of significant performance characteristics of commonly used lamps. Explanation of the parameters will be provided later in this chapter. Figure 6-4 shows shapes of commonly available lamps. Each lamp shape also includes the corresponding American National Standards Institute (ANSI) designation used by many lamp manufacturers in their catalogs. The designation is typically followed by a number, which expresses the diameter of the lamp in multiples of 1/8 inch, so that T-12 refers to a tubular fluorescent lamp with a diameter of 12/8 or 1.5 in. (38 mm), and PAR 30 is a parabolic reflector lamp with a diameter of 30/8 or 3.75 in. (95 mm). Manufacturers use a variety of bases, discussed in the individual lamp sections below. Light sources, luminaires, controls, and system layout are closely interrelated. A light source that is appropriate for one type of application may be impractical for another. See the application chapters that follow for guidance on light source selection.

INCANDESCENT FILAMENT AND TUNGSTEN-HALOGEN LAMPS The primary consideration of filament lamp design is that it will produce the spectral radiation desired (visible, infrared, ultraviolet) most economically for the application intended. Realization of this objective in an incandescent filament lamp requires the specification of the following: filament material, length, diameter, form, coil spacing, and mandrel size (the mandrel is the form on which the filament is wound); lead-in wires; number of filament supports; filament mounting method; vacuum or filling gas; gas pressure; gas composition; and bulb size, shape, glass composition, and finish. The manufacture of high-quality lamps requires adherence to these specifications and necessitates careful process controls.

Figure 6-1. Time line of light sources. The construction and principle of operation of incandescent filament and tungsten-halogen lamps are similar; however, the halogen regenerative cycle enables a tungsten-halogen lamp to provide the following benefits compared to a conventional incandescent lamp: longer life, higher color temperature, higher efficacy, and no bulb blackening.

Incandescent Lamp Construction Filaments. The efficacy of light production depends on the temperature of the filament. The higher the temperature of the filament, the greater the portion of the radiated energy that falls in the visible region. For this reason it is important in the design of a lamp to keep the filament temperature as high as is consistent with satisfactory life. For example, iron is not a good filament material because it melts at a relatively low temperature (1527°C) for efficient light production. Numerous materials have been tested for filament suitability. Desirable properties of filament materials are a high melting point, low vapor pressure, high strength, high ductility, and suitable radiating characteristics and electrical resistance. Tungsten for Filaments. Early incandescent lamps used carbon, osmium, and tantalum filaments, but tungsten has many desirable properties for use as an incandescent light source. Its low vapor pressure and high melting point, 3382°C (6120°F), permit high operating temperatures and consequently high efficacies. Drawn tungsten wire has high strength and ductility, allowing the uniformity necessary for present-day lamps. Alloys of tungsten with other metals such as rhenium are useful in some lamp designs. Thoriated tungsten wire is used in filaments for rough service applications. Radiating Characteristics of Tungsten.4-7 The ratio of the radiant exitance of a thermal radiator to that of a blackbody radiator is called the emissivity, and thus the emissivity of a blackbody is 1.0 for all wavelengths (see Chapter 1, Light and Optics). Tungsten is a selective radiator because its emissivity is a function of the wavelength. Figure 6-5 illustrates the radiation characteristics of tungsten and of a blackbody and shows that for the same amount of visible radiation, tungsten radiates only a percentage of the total radiation from a blackbody at the same temperature. (The intensity of curve B is approximately 76% of curve A.) Only a small percentage of the total radiation from an incandescent source is in the visible region of the spectrum. As the temperature of a tungsten filament is raised, the radiation in the visible region increases (Figure 6-6), and thus the luminous efficacy increases. The luminous efficacy of an uncoiled tungsten wire at its melting point is approximately 53 lm/W. In order to obtain long life, it is necessary to operate a filament at a temperature well below the melting point, resulting in a loss in efficacy. Resistance Characteristics of Tungsten. Tungsten has a positive resistance characteristic, so that its resistance at operating temperature is much greater than its cold resistance. In general-service lamps, the hot resistance is 12 to 16 times the cold resistance. Figure 6-7 illustrates the change in resistance of the tungsten filament with temperature for various lamps. The low cold resistance of tungsten filaments results in an initial in-rush of current that, because of the reactive impedance characteristic of the circuit, does not reach the theoretical value indicated by the ratio of the hot-to-cold resistance. Figure 6-8 gives the effect of the change in resistance on the current in incandescent filament lamps. The in-rush current due to incandescent filament loads is important in the design and adjustment of circuit breakers, in circuit fusing, in the design of lighting-circuit switch contacts, and in dimmer designs.

Figure 6-2. Continued

Figure 6-2. Approximate Luminance of Various Light Sources Color Temperature. Often it is important to know the apparent color temperature of an incandescent lamp. Figure 6-9 expresses the approximate relationship between color temperature and luminous efficacy for a range of gas-filled lamps. The efficacy value often can be found in the literature, or it can be calculated from published lumen and wattage data. From this value it is possible to approximate the average color temperature of the filament. Construction and Assembly. Figure 6-10 shows the basic parts and steps in the assembly of a typical incandescent, general-service filament lamp. In miniature lamps three methods of construction are typically used: flange seal, butt seal, and pinch seal (Figure 6-11). The flange seal generally is used with lamps 20 mm (0.79 in.) and larger in bulb diameter. This construction features a glass stem with a flange at the bottom that is sealed to the neck of the bulb. When used with bayonet bases, the plane of the filament and lead wires is normally at right angles to the plane of the base pins, but a tolerance of 15°generally is permitted. The advantages of this construction are: (1) heavy lead-in wires can be used for lamp currents up to 12 A, (2) the filament can be accurately positioned, and (3) sturdy stem construction resists filament displacement and damage from shock and vibration. The butt seal is constructed as follows. A mount consisting of lead-in wires, bead, and filament is dropped into the open end of the bulb. The lead-in wires are bent to locate the filament at the desired distance from the bulb end. An exhaust tube is then dropped down and butted against the lead-in wire and glass bulb just prior to sealing and exhausting. The base, applied later, together with the basing cement, must not only provide the lamp contacts but also protect the delicate seal. Because of seal limitations, butt seal lamps are restricted to small wire sizes with a current limit of approximately 1.0 A. The filament position

varies considerably more than in flange seal lamps, since there is no definite relationship between the planes of the filament and base pins. Occasionally butt seal lamps are used without bases; these lamps should be handled carefully. When used with a base, the advantages of butt seal construction are (1) low cost and (2) small size (usually 20 mm [0.79 in.] and below). The pinch seal is so named because glass is pinched, or formed, around the lead-in wires. Two forms are used: wire terminals and wedge base construction. For the smaller types of glow lamps, the bulb is exhausted and tipped off at the end opposite the lead-in wires. With newer wedge base lamps, the exhaust tip is at the bottom rather than the top. Pinch seal construction eliminates the need for a conventional base. Advantages are: (1) low cost; (2) small size; (3) with filament lamps, the elimination of solder and cement, which allows operation up to 300°C; and (4) small space required for wedge base lamps. The molybdenum seal is used in some tungsten-halogen lamps, where wire seals cannot be employed, due to their thermal expansion mismatch with the fused silica envelope material. Molybdenum seals consist of thin ribbons or foils of molybdenum that are pinched in the base of the lamps to provide the required electrical lead-in. The ribbons provide a reliable seal as long as base temperature is kept below the molybdenum oxidation temperature of 350°C (662°F). Filament Forms and Designations. Filament design involves a careful balance between light output and life. Filament forms, sizes, and support constructions vary widely with different types of lamps (Figure 6-12). Their designs are determined largely by service requirements. Filament forms are designated by a letter or letters followed by an arbitrary number. The most commonly used letters are: S (straight), meaning the wire is uncoiled; C (coiled), meaning the wire is wound into a helical coil; and CC (coiled coil), meaning the coil is itself wound into a helical coil. Coiling the filament increases its luminous efficacy; forming a coiled coil further increases efficacy. More filament supports are required in lamps designed for rough service and vibration service than for general-service lamps. Bulbs Shapes and Sizes. Common bulb shapes are shown in Figure 6-4. Types of Glass. Most bulbs are made of regular lead or soda lime (soft) glass, but some are made of borosilicate heat-resisting (hard) glass. The latter withstand higher temperatures and are used for highly loaded lamps. They usually withstand exposure to moisture or luminaire parts touching the bulb. Three specialized forms of glass are also used as lamp envelopes: fused silica (quartz), high-silica, and aluminosilicate glass. These materials can withstand still higher temperatures. See the section "Bulb and Socket Temperature" below.

Figure 6-3. General Characteristics of Commonly Used Light Sources* (This table is intended to show the wide range of parameters available for lamp products. A specific example has been chosen for each source type.)

Figure 6-4. Typical bulb shapes (not to scale) and their ANSI designations. Not every ANSI designation, as key-listed here to a descriptive phrase or word, is illustrated.

Bulb Finishes and Colors. Inside frosting is applied to many types and sizes of bulbs. It produces moderate diffusion of the light with very little reduction in output. The extremely high filament luminance of clear lamps is reduced, and striations and shadows are mostly eliminated. White lamps having an inside coating of finely powdered white silica provide a better diffusion with little absorption of light. Daylight lamps have bluish glass bulbs that absorb some of the long wavelengths produced by the filament. The transmitted light is of a higher correlated color temperature. This color, achieved at the expense of approximately 35% reduction in light output through absorption, varies between 3500 and 4000 K. This is almost midway between tungsten filament light and daylight.

Figure 6-5. Radiating characteristics of tungsten. Curve A: radiant flux from one square centimeter of a blackbody at 3000 K. Curve B: radiant flux from one square centimeter of tungsten at 3000 K. Curve B′: radiant flux from 2.27 square centimeters of tungsten at 3000 K (equal to curve A in visible region). (The 500-watt 120-volt general service lamp operates at about 3000 K.)

Figure 6-6. Spectral power distribution in the visible region from tungsten filaments of equal wattage but different temperatures.

Figure 6-7. Variation of tungsten filament hot resistance with temperature.

Figure 6-8. Effect of Hot-Cold Resistance on In-rush Current in an Incandescent Filament (Laboratory Conditions)

Figure 6-9. Variation of color temperature with lamp efficacy. General-service incandescent colored lamps are available with inside- and outside-spray-coated, outside-ceramic, transparent-plastic-coated, and natural-colored bulbs. Outside-spray-coated lamps generally are used indoors and not exposed to weather. Their surfaces collect dirt readily and are not easily cleaned. Insidecoated bulbs have smooth outside surfaces that are easily cleaned; thus the pigments are more durable. Ceramic-coated bulbs have the colored pigments fused onto the glass, providing a permanent finish. They are suitable for indoor and outdoor use, as are most transparent-plastic-coated bulbs. The coating permits the filament to be observed directly. Natural-colored bulbs are made of colored glass. Colored reflector lamps use ceramic-coated bulbs, stained bulbs, plasticcoated bulbs, and dichroic interference filters to obtain the desired color characteristics. Bases. Figure 6-13 shows the most common lamp bases. Most lamps for general lighting purposes employ one of the screw bases. Where a high degree of accuracy in positioning of light sources with relation to optical elements is important, as in the case of projection systems, bipost and prefocus bases ensure proper filament location. Lamp wattage is also a factor in determining the base type. Most bases are secured to the bulbs by cement and are cured by heat when the lamp is manufactured. Since this cement becomes weaker with age, particularly if exposed to excessive heat, lamps intended for high-temperature service use a special heat-tolerant basing cement or bases that are mechanically fastened without the use of cement. Gas Fill. Around 1911, attempts were made to reduce the rate of evaporation of the filament by the use of gas-filled bulbs. Nitrogen was first used for this purpose.8 Although the fill gas reduced bulb-wall blackening, it increased heat loss, leading to even greater light loss. An incandescent filament operating in an inert gas is surrounded by a thin sheath of heated gas, to which some of the input energy is lost; the proportion lost decreases as the filament diameter is increased. When the filament is coiled in a tight helix, the sheath surrounds the entire coil so that the heat loss is no longer determined by the diameter of the wire but by the diameter of the coil, thus greatly reducing this energy loss. A coiled-coil filament has even less length for a given power rating, thus further reducing the area available for convective cooling. The use of coiled-coil filaments and gas-filled bulbs has yielded major improvement in incandescent lamp efficacies. However, general-service 120-V lamps below 25 W are usually of the vacuum type since gas filling does not improve the luminous efficacy in this wattage range.

Figure 6-10. Steps in the manufacture of a typical incandescent filament lamp.

Figure 6-11. Primary type of bulb construction: (a) glass halogen capsule with pinched seal, (b) pinch seal with lead-in wire terminals, (c) butt seal, (d) flange seal, and (e) molybdenum ribbon pinch seal. Inert gases are now preferred because they do not react with the internal parts of the lamp and because they conduct less heat than nitrogen. It was some years after the development of gas-filled lamps before argon became available in sufficient quantity and purity and at reasonable cost. Most lamps are now filled with argon and a small amount of nitrogen; some nitrogen is necessary to suppress arcing between the lead-in wires. The proportion of argon and nitrogen depends on the voltage rating, the filament construction and temperature, and the lead-tip spacing. Typical amounts of argon in use are: 99.6% for 6-V lamps, 95% for 120-V general-service coiled-coil lamps, 90% for 230-V lamps having fused lead wires, and 50% or less for 230-V lamps when no fuses are used in the leads. Some projection lamps are 100% nitrogen filled.

Figure 6-12. Typical lamp filament constructions (not to scale). Krypton, although expensive, is used in some lamps where the increase in cost is justified by the increased efficacy or increased life. Krypton gas has lower heat conductivity than argon. Also, the krypton molecule is larger than that of argon and therefore further retards the evaporation of the filament. Depending on the filament form, bulb size, and mixture of nitrogen and argon, krypton fill can increase efficacy by 7 to 20%.9,10 Krypton is used in some special lamps such as marine signal and miner's cap lamps because of the resulting high efficacy. An even more expensive gas, xenon, is used in a number of unique applications. Xenon has lower heat conductivity and larger molecular size than krypton and thus allows even higher efficacy. However, it costs significantly more than krypton and is used only for special product applications. Hydrogen gas has a high heat conductivity and is therefore useful for signaling lamps where quick flashing is desired.11 Tungsten-Halogen Lamps. The light generation mechanism of tungsten-halogen lamps is the same as that of common incandescent filament lamps, except for the halogen regenerative cycle. Halogen is the name given to a family of electronegative elements, including bromine, chlorine, fluorine, and iodine. Although the tungsten-halogen regenerative cycle has been understood for many years, no practical method of using it was established until the development of smalldiameter fused quartz envelopes for filament lamps provided the proper temperature parameters. Iodine was used in the first tungsten-halogen lamp; today, other halogen compounds, predominantly bromine, are used. The regenerative cycle starts with the tungsten filament operating at incandescence, evaporating tungsten off the filament. Normally the tungsten particles would collect on the bulb wall (Figure 6-14a), resulting in bulb blackening, common with incandescent lamps and most evident near the end of their life. However, in halogen lamps the temperature of the bulb is high enough so that the tungsten combines with the halogen. The correlated minimum temperature of the bulb must be approximately 260°C (500°F). The resulting tungsten-halogen compound is also gaseous and continues to circulate inside the lamp until it comes in contact with the incandescent filament. Here, the heat is sufficient to break down the compound into tungsten, which is redeposited on the filament, and halogen, which is freed to continue its role in the regenerative cycle (Figure 6-14b). However, since the tungsten does not necessarily redeposit exactly where it came from, the tungsten-halogen lamp still has a finite life. Dimmed tungsten-halogen lamps should periodically be run at full power, inducing the tungsten-halogen cycle to clean the tungsten off the bulb wall, and thereby maintaining lamp efficacy over time.

Figure 6-13. Common lamp bases (not to scale). ANSI designations are shown, where available. The requirement of a high bulb wall temperature for the halogen regenerative cycle led to the corollary effect of producing smaller lamps. This resulted first in the development of small low-voltage reflector lamps and finally in the incorporation of halogen capsules in various PAR envelopes. These PAR envelopes have replaced both incandescent PAR and BR types.

Figure 6-14. Tungsten cycle in (a) standard incandescent, and (b) tungsten-halogen.

Figure 6-15. Luminous and Thermal Characteristics of Typical Incandescent Filament Lamps

Performance Parameters Energy Characteristics. The manner in which the energy input to a lamp is dissipated can be seen in Figure 6-15 below for typical general-service lamps. The radiation in the visible spectrum (column 2) is the percentage of the input power actually converted to visible radiation. The gas loss (column 4) indicates the amount of heat lost by the filament due to the conduction through and convection by the surrounding gas in gas-filled lamps. The end loss (column 6) is the heat lost from the filament by the lead-in wires and support hooks which conduct heat from the filament. Column 3 shows the total radiation beyond the bulb, which is less than the actual filament radiation due to absorption by the glass bulb and the lamp base. Incandescent lamps operated below 25 Hz will produce perceptible flicker and can create a stroboscopic effect. Flicker will be less from an incandescent source if it has a larger filament and is operated at a higher wattage and at a higher supply frequency. Modern incandescent light sources operated at 60 Hz do not produce noticeable flicker, nor a stroboscopic effect, to the human eye. The flicker index of several incandescent lamps operated at 25 Hz and 60 Hz is shown in Figure 6-16. Consult Chapter 2, Measurement of Light and Other Radiant Energy, for more information about lamp flicker.

Figure 6-16. Flicker Characteristics of Incandescent Filament Lamps Bulb and Socket Temperatures. Incandescent filament lamp operating temperatures are important for several reasons. Excessive lamp temperature can affect lamp life, luminaire life, and the life of the electrical supply circuit. High temperatures can ignite combustible materials that form a part of the luminaire or those adjacent to the luminaire. Under certain atmosphere or dust conditions, high bulb temperatures (above 160°C [320°F]) can induce explosion or fire. Bulb and socket temperatures for a 100-W A-19 lamp and a 500-W PS-35 lamp for different operating positions are shown in Figure 6-17. General-service incandescent filament lamp bulbs are made of regular lead or lime soft glass, the maximum safe operating temperature of which is 370°C (698° F). Some lamps for special applications, such as outdoor floodlighting lamps, have hard glass bulbs that have a safe temperature limit of 470°C (878°F). Lamps with still harder glass bulbs can be operated up to 520°C (968°F). The maximum safe base temperature for general-service lamps is 170°C (338°F), measured at the junction of the base and the bulb. In all cases excessive temperature can cause failure of the basing cement, as well as softening of the solder used to connect the lead wires to the base. Silicone cement and highmelting-point solder permit base temperatures to approach 260°C (500°F). Bipost bases carry considerable heat to the socket through the base pins, and the parts of the socket in contact with the base pins should be capable of withstanding temperatures up to 290°C (554°F). Tubular fused quartz infrared and tungstenhalogen cycle lamps generally have a maximum seal temperature of 350°C (662°F) to prevent oxidation.

Figure 6-17. Incandescent filament lamp operating temperatures in still air at 25°C (77°F) ambient: (a) 100-watt CC -8, A-19 lamp; (b) 500-watt CC-8, PS-35 lamp. All lamp temperatures shown are in degrees Celsius. Some of the factors affecting base temperature are filament type, light center length, heat shields, bulb shape and size, and gas fill. Base temperature is not necessarily correlated to wattage ratings; that is, lamps of lower wattage do not necessarily have lower base temperatures. Medium-base luminaires should be capable of accepting lamp base temperatures on the order of 135-150°C (1275-1302°F). These limits are consistent with Canadian Standards Association (CSA), American National Standards Institute (ANSI), and International Electrotechnical Commission (IEC) standards for base temperature for A-type lamps. Measurements should be made in accordance with ANSI C78.25, "Lamp-Base Temperature Rise-Method of Measurements."12 If the luminaire accepts R and PAR lamps, then the temperature capability of the luminaire should be 170 to 185°C (338 to 365°F) if reasonable electrical insulation life is desired for the luminaire. Heat-transmitting dichroic-reflector lamps should be placed only in luminaires specifically designed for them; this is normally indicated on the luminaire. For base-up operation, and where only slight enclosing of the lamp is provided by the luminaire, base temperature is the major factor affecting luminaire temperatures, with lamp wattage being a minor consideration. As the luminaire provides more and more enclosing of the lamp, base temperature has less effect, and wattage assumes more importance. High temperatures reduce the life of electrical insulation for lamp and luminaire parts. Figure 6-18 shows how the position of the filament affects the temperature of the bulb. This variation in temperature with position is sufficiently large to affect the life of wire insulation and other luminaire components. Lamp Characteristics Life, Efficacy, Color Temperature, and Voltage Relationships. If the voltage applied to an incandescent filament lamp is varied, there is a resulting change in the filament resistance and temperature, current, power, light output, efficacy, and life. These characteristics are interrelated, and not one of them can be changed without affecting the others. The following equations can be used to calculate the effect of a change from the design conditions on lamp performance (capital letters represent normal rated values; lowercase letters represent changed values):

Figure 6-18. Profile temperatures vary with rotation of lamp filament; three types shown.

For approximations, the following exponents may be used in the above equations: d = 13, g = 1.9, k = 3.4, n = 1.6, and m = 0.42. For more accuracy, the exponents must be determined by each lamp manufacturer from a comparison of normal-voltage and over- or undervoltage tests of many lamp groups. Exponents vary for different lamp types, lamp wattages, and ranges of percentage voltage variation. The values given above are roughly applicable to vacuum lamps of approximately 10 lm/W and gas-filled lamps of approximately 16 lm/W in a voltage range of 90 to 110% of rated voltage. For information outside this

range, refer to Figure 6-19. The curves of Figure 6-19a show the effect of voltage variations on lamps in general lighting (multiple) circuits.13,14 The effect of voltage variation on the characteristics of tungsten-halogen lamps cannot be accurately predicted outside of the voltage range of 90 to 110% of the rated voltage. Filament Notching. Ordinarily, for laboratory test operation, normal tungsten filament evaporation determines incandescent lamp life. Where that is so, lamps should reach their design-predicted life. Another prominent factor influencing filament life is filament notching. Filament notching is the appearance of steplike or sawtooth irregularities on all or part of the tungsten filament surface after long use. These notches reduce the filament wire diameter at random points. In some cases, especially for fine-wire filaments, the notching is so deep as to almost sever the wire. Faster spot evaporation due to high temperatures at this notch and reduced filament strength become the dominant factors influencing lamp life. Predicted lamp life can be reduced by as much as one-half from this cause. Filament notching is associated with at least three factors (primarily occurring in fine-wire filament lamps): (1) low-temperature filament operation, as in longlife lamps with 10,000- to 100,000-h designs; (2) small filament wire sizes, typically less than 0.025 mm (0.001 in.) in diameter; and (3) direct current operation. Depreciation During Life. Over a period of time, incandescent filaments evaporate and become smaller, which increases their resistance. In multiple circuits, the increase in filament resistance reduces current, power, and light. A further reduction in light output is caused by the absorption of light by the deposit of the evaporated tungsten particles on the bulb. In series circuits having constant-current regulators, the increase in filament resistance during life causes an increase in the voltage across the lamp and a consequent increase in wattage and generated lumens. This increase in lumens is offset to varying degrees by the absorption of light by the tungsten deposit on the bulb. In low-current lamps the net depreciation in light output during life is very small, or in the smaller sizes there may be an actual increase. In lamps of 15- and 20-A ratings, the bulb blackening is much greater, and throughout life more than offsets the increase in lumens due to the increased wattage.

Figure 6-19. Effect of voltage and current variation on the operating characteristics of: (a) incandescent filament lamps in general lighting (multiple) circuits and (b) tungsten-halogen lamps in series street lighting circuits. The blackening in vacuum lamps is uniform over the bulb. In gas-filled lamps the evaporated tungsten particles are carried by convection currents to the upper part of the bulb. When gas-filled lamps are burned base up, most of the blackening occurs on the neck area, where some of the light is normally intercepted by the base. Consequently, the lumen maintenance for base-up operating is better than for base-down or horizontal operating with gas-filled lamps. In a base-up operating lamp, an appreciable reduction of lamp lumen depreciation (LLD) can be obtained through the use of a coiled-coil filament located on or parallel to the bulb axis. To reduce blackening from traces of oxygen or water vapor in the gas fill, an active chemical, known as a getter, is used inside the bulb to combine with and absorb impurities remaining in the bulb. Tungsten-halogen cycle lamps generally have significantly less depreciation during life due to the regenerative cycle, which removes the evaporated tungsten from the bulb and redeposits it on the filament. Figure 6-20 shows the change in light output and efficacy for typical incandescent and tungsten-halogen lamps. See the discussion on tungsten-halogen lamps earlier in this chapter.

Figure 6-20. Typical operating characteristics of lamps as a function of burning time: (a) general-service lamps and (b) tungsten-halogen lamps. (Note differences in scales.) Lamp Mortality. Many factors inherent in the manufacturing process make it impossible for every lamp to achieve the rated life for which it was designed. For this reason, lamp life is rated as the average of a large group. A range of typical mortality curves representing the performance of high-quality lamps is illustrated in Figure 6-21. Dimming of Incandescent and Tungsten-Halogen Lamps. Dimmers today have a dual purpose: energy conservation and aesthetic lighting effects. Incandescent lamps can be dimmed simply by lowering the voltage across the lamp filament. When the voltage is lowered, less power is dissipated and less light is produced with a lower color temperature. An added benefit is an increase in the life of an incandescent lamp. For example, when an incandescent lamp is operated at 80% rated voltage using a dimmer, its life is increased by a factor of nearly 20. In a tungsten-halogen lamp, the life of the filament depends on voltage just as with standard incandescent filament lamps. However, because the regenerative halogen cycle stops when bulb wall temperature falls below 260° C (500°F), the tungsten halogen lamp blackens and its useful life is not extended by nearly the same factor as that of standard lamps. This can be partially compensated for by periodically operating the lamp near or at full light output, which helps clean the lamp of tungsten deposits. In the 1950s, rheostats were used for dimming by regulating the lamp current in an incandescent lamp. They were large and inefficient. Today, most dimmers are electronic, using thyristor and transistor circuits that have low power dissipation. Modern dimmers are efficient and reduce power as the source is dimmed. Thyristors operate as high-speed switches that rapidly turn the voltage to the lamp on and off. This switching can cause electromagnetic interference with other electrical equipment as well as audible buzzing in the lamp filament. Magnetic coils known as chokes are usually used as filters to reduce these effects. With many wall-box dimmers, however, lamp buzzing cannot be completely eliminated because a larger choke is needed than space allows. For these cases, remotely mounted, properly sized lamp debuzzing coils or additional chokes are recommended.

Figure 6-21. Range of typical mortality curves (average for a statistically large group of incandescent filament lamps).

Classification of Incandescent and Tungsten-Halogen Lamps Incandescent filament lamps were historically divided into three major groups: large lamps, miniature lamps, and specialty lamps. They also were cataloged separately by lamp manufacturers. There is no sharp dividing line among the groups, and they are usually included in the same catalog by the manufacturer. The discussion below classifies lamps into four categories for the purposes of illustrating their applications: general lamps, dedicated application lamps, low-voltage miniature and sealed beam lamps, and photographic and photo-optical lamps. The following gives a brief description of a few of the many types of lamps that are regularly manufactured. More complete details are available in manufacturers' catalogs. General Lamps (Lamps Used in a Variety of Applications) General Service. These are large lamps made for general lighting use on 120-V circuits. General-service lamps range from 10 to 1500 W and satisfy the majority of incandescent lighting applications. All sizes are made in both clear and inside-frosted bulbs. Below 200 W, inside-white-coated finishes are also available. High Voltage. This voltage class refers to lamps that operate directly on circuits of 220 to 300 V and represents a very small portion of the lamp demand in North America. Lampholders should be Underwriters Laboratories (UL) or CSA approved for a voltage level appropriate to the voltage rating of the lamps being used, that is, 250 V for lamps up to 250 V, 600 V for lamps above 250 V. High-voltage lamps have filaments of small diameter and longer length and require more supports than corresponding 120-V lamps. Therefore, they are less rugged and are 25 to 30% less efficient because of greater heat losses. Due to the higher operating voltage, these lamps require less current for the same power, permitting some wiring economy.15 Extended Service. Extended-service lamps are intended for use in applications where a lamp failure causes significant inconvenience, nuisance, or hazard, or

where replacement labor cost is high or power cost is unusually low. For such applications, where long life is most important and a reduction of light output is acceptable, lamps with 2500-h or longer rated life are available. Longer life is obtained by operating the filament at lower temperatures than normal. This, however, lowers the luminous efficacy. In most general-service use the cost of power used during the lamp life is many times the lamp life cost, and therefore efficacy is important. Where replacement of burned-out lamps is an easy, convenient operation, as in residential use, long-life lamps usually are not recommended.16 For such use, incandescent lamps with the usual 750- or 1000-h design life give a lower cost of light than extended-service lamps. General-Lighting Tungsten-Halogen Lamps. These lamps improve on the regular incandescent sources in several ways. Their advantages over regular incandescent lamps include low LLD and compactness. They also provide whiter light (higher color temperature) and longer life at a given light output. Halogen lamps are available in both line voltage and low voltage designs. The line voltage products are available as single ended, double ended, and PARs. The low voltage types are generally capsules or small reflector types. For specialty applications, other voltage ratings (e.g., 84 V, 200 V) are also available. Low voltage halogen lamps operate in the voltage range from 5 to 30 V. This voltage is supplied through a step-down transformer. The advantages of lowvoltage lamps are greater resistance to vibration and shock because of their larger diameter filament wire, a more compact filament that allows better beam control, and higher efficacy than line voltage lamps. There is more ultraviolet (UV) radiation generated from tungsten-halogen lamps than from regular incandescent lamps, due to the higher filament temperature. The amount of UV radiation emitted is determined by the envelope material. Fused quartz and most high-silica glass transmit most of the UV radiated by the filament, while other special high-silica glass and aluminosilicate glass absorb UV radiation. In general lighting applications, luminaires for tungsten-halogen lamps should have a lens or cover glass that, in addition to providing the required safety protection in case of lamp breakage, filters out most of the UV radiation. In applications where the reduction of UV radiation is critical, additional filtering might be required. Care should be taken when applying lamps operating at correlated color temperatures above 3100 K, since both ultraviolet and short-wavelength visible radiation increase with color temperature,17 creating potential hazard for both people and objects. The halogen regenerative cycle is very temperature sensitive. Operating lamps at voltages above or below manufacturer's recommendations can have adverse effects on the internal chemical processes. It is also important to follow manufacturer's recommendations as to operating position, bulb handling, and luminaire temperatures. The higher pressures used allow tungsten-halogen lamps to be designed for higher efficacy and longer life than normal incandescent lamps of the same wattage. For example, the 500-W nonregenerative cycle lamp is rated at 10,600 lm for 1000 h, while the 500-W T-3 tungsten-halogen lamp is rated at 10,950 lm for 2000 h. The development of high-performance glass tubings has led to the use of halogen glass capsules in PAR lamps, including PAR 16, PAR 20, PAR 30, and PAR 38. Multilayer infrared reflective coatings can be applied to both quartz and glass halogen bulbs. These coatings transmit light and reflect infrared energy back to the filament. This reduces the input power required to reach a given filament temperature. Luminous efficacies of 27.0 to 35.6 lm/W are obtained with this technique without reducing life. Reflector. Reflector lamps include those made in standard and special bulb shapes and have a reflective coating applied directly to part of the bulb surface. Both silver and aluminum coatings are used.18 Silver coatings can be applied internally or externally, and in the latter case the coating is protected by an electrolytically applied copper coating and sprayed aluminum finish. Aluminum coatings are applied internally by condensation of vaporized aluminum on the bulb surface. The following reflector lamps are readily available: bowl reflector lamps, neck reflector lamps, reflector spot and reflector flood lamps in R-type bulbs (certain sizes of reflector lamps are available with heat-resisting glass bulbs), and ellipsoidal reflector lamps (ER types), which allow substantially improved efficiency in deep, well-shielded downlights.19 PAR spot and PAR flood lamps use PAR bulbs, typically constructed from two molded glass parts, the reflector and the lens, which are fused together. As the names suggest, several lamp designs with different candlepower distributions are available, typically expressed in terms of beam angle. Colored R and PAR lamps are available. Cool beam PAR lamps with heat-transmitting dichroic reflectors are available where it is desirable to reduce the infrared energy in the beam. Long-neck halogen PAR-30 lamps are available for retrofitting standard R-30 luminaires without the use of socket extenders. Multifaceted pressed glass reflector lamps with tungsten-halogen capsules and infrared-transmitting dichroic reflectors, known as MR11 and MR-16, have been adapted from projection lamp designs for display lighting applications. Rough Service. To provide the resistance to filament breakage, rough-service lamps employ special, multiple-filament-support construction (C-22 in Figure 612). Because of the number of supports, the heat loss is higher and the efficacy lower than for general-service lamps. In using miniature lamps where rough-service conditions are encountered, bayonet and wedge base lamps should be chosen instead of screw base lamps. Bayonet and wedge base lamps lock in the socket; screw base lamps tend to work loose. Linear Incandescent. The linear incandescent lamp has a tubular bulb diameter of 26 mm (1 in.) and two metal disk bases, one at each end of the lamp, with the filament connected between them. Many have adjustable insert tabs attached to the bases, thereby simplifying insertion into their sockets. The filament, in the form of a stretched coil, is supported on glass insulating beads along a small metal channel within the bulb. The 30- and 60-W sizes are available in the 450-mm (18-in.) length. The 40-W lamp is made in a 300-mm (12-in.) length. All sizes are available in either clear or inside-frosted tubes as well as white and various color coatings.

Figure 6-22. The beam angle is the angle within which the lamp produces 50% of the lamp's maximum intensity. Another style of linear incandescent lamp, using the S14 base, is available. Versions include 35, 60, and 150 W. Decorative. A wide variety of lamps for decorative application is available. Different bulb shapes, together with numerous colors and finishes, are used to achieve the desired appearance. Lamp manufacturers' catalogs should be consulted for information on the many decorative types. Three-Way. These lamps employ two filaments, operated separately and in combination, to provide three light levels. The common lead-in wire is connected to the shell of the base; the other end of one filament is connected to a ring contact, and the end of the other filament to a center contact. Thus, either filament, or both together, may be used with the selection made in the socket. Three-way lamps are available in several different wattage combinations. Energy-Saving. Most general- and extended-service lamps are now available in reduced wattages, which also have lower lumen output. The use of reflector lamps to improve directional lighting has also become popular for energy-saving applications. Better optical control with sharper cutoffs places the same illuminance on a specific area with fewer watts. A variety of reflector lamps is now available, including the BR30 and BR40. The most significant energy improvement is in the family of halogen PAR and halogen IR PAR products, which use significantly less power and provide equal or superior illuminance than incandescent PAR and BR types. In traffic signals, lamps are now being marketed with a ring mirror so as to capture lost light and redirect it through the lens. Dedicated Application Lamps. Dedicated application lamps are those that have been designed primarily for a single application. Spotlight and Floodlight. Lamps used in spotlights, floodlights, and other specialized luminaires for lighting theater stages, motion picture studios, and television studios, have concentrated filaments accurately positioned with respect to the base. When the filament is placed at the focal point of a reflector or lens system, a precisely controlled beam is obtained. These lamps are intended for use with external reflector systems. Because of their construction, these lamps must be burned in positions for which they are designed, to avoid premature failures. Appliance Lamps. High-temperature types that can withstand temperatures up to 315°C (599°F) are available for ovens. Other types are designed for refrigerators, and rugged filament types are used in vacuum cleaners. Vibration Service. Most lamps have coiled filaments made of tungsten with high sag resistance. Vibration lamps, designed for use where vibrations would cause early failure, are made with a more malleable tungsten filament. The sagging of the wire used allows the coils to open up under vibration, thus preventing short circuits between coils. To withstand shock and vibration, low-voltage (6.3-V or less) miniature radio panel lamps incorporate mounts whose resonant frequency has been synchronized with that of the coiled filament. Sometimes only trial and error will determine the best lamp to resist shock and vibration. Vibration-resistant sockets or equipment, using a coiled spring or other flexible material to deaden vibration, have been employed where general-service lamps are used under conditions of severe vibration. Showcase. Display cases in retail applications commonly use tubular bulbs with conventional screw bases. The longer lamps have filaments with supports similar to linear incandescent lamps. The common sizes are 25 and 40 W, but sizes up to 75 W are available. Sign. While large numbers of gas-filled lamps are used in enclosed and other types of electric signs, those designated particularly as sign lamps are mostly of the vacuum type. These lamps are best adapted for exposed sign and festoon service because the lower bulb temperature of vacuum lamps minimizes the occurrence

of thermal cracks resulting from rain and snow. Some low-wattage lamps, however, are gas filled for use in flashing signs. Bulb temperatures of these lowwattage, gas-filled lamps are sufficiently low to permit exposed outdoor use on high-speed flashing circuits. Traffic Signal Lamps. Lamps used in traffic signals are subjected to more severe service requirements than in most applications of incandescent lamps. Such lamps must be compatible with the design requirements of standard traffic signals. Ribbon Filament Lamps. Incandescent lamps made with ribbons or strips of tungsten for the filaments have been used in special applications where it is desirable to have a substantial area of fairly uniform luminance.20 Ribbon dimensions vary from 0.7 to 4 mm (0.03 to 0.16 in.) in width and up to 50 mm (1.97 in.) in length. The 5-20-A ribbon filament lamps are usually employed in recorders, instruments, oscillographs, and microscope illuminators. The 30- to 75-A lamps are used for pyrometer calibration standards and for spectrographic work. Figure 6-23 shows typical lamps.

Figure 6-23. Typical ribbon filament lamps: (a) 6-volt, 18-ampere, T-10, 2-mm, 3000-K microscope illuminator; (b) 6-volt, 9-ampere, T-8—½, 1mm, 3000-K optical source; and (c) 3.5-volt, 30-ampere, T-24 with quartz window, 3-mm, U-shaped filament, 2300-K pyrometer and spectroscope source. Infrared Heat Lamps. All incandescent filament lamps are effective generators of infrared (IR) radiation. Most of the incandescent filament lamp input power is radiated as IR energy. Wavelengths longer than 5000 nm are absorbed to a large extent by the glass or fused quartz envelope. Lamps for heating applications are designed for low light output and long life. Tubular fused quartz lamps are also available with a ceramic reflector that increases heat by approximately 50% directly below the center line of the lamp. Tungsten filament IR lamps are available with ratings up to 5000 W. Generally speaking, tungsten filament lamps for industrial, commercial, and residential service operate at a filament color temperature of 2500 K. At this low operating temperature compared to lighting lamp filament temperatures, the service life is well in excess of 5000 h. Frequently, lamps using tungsten filaments have provided many years of operation because the service life is generally determined by mechanical breakage or rupture of the filament due to vibration or handling, rather than the rate of evaporation of tungsten, as is the case with lamps used for light. Lamps having heat-resisting glass bulbs or tubular fused quartz envelopes are recommended where liquids might come in contact with the bulb. The distribution of power radiated by various infrared sources is shown in Figure 6-24. For application information, see Chapter 5, Nonvisual Effects of Optical Radiation. Low-Voltage Miniature and Sealed Beam Lamps. The term "miniature" applied to light sources is a lamp manufacturer's designation determined by the trade channels through which these lamps are distributed, rather than by the size or characteristics of the lamps. In general, however, most miniature lamps are small and require relatively little power. The most notable exceptions to this generalization are sealed beam lamps, such as automotive headlamps and aircraft landing lamps, some of which are classed as miniature lamps, even though they can be as large as 200 mm (7.3 in.) in diameter and dissipate up to 1000 W.

Figure 6-24. Spectral power distribution from various infrared sources. The great majority of miniature lamps are either incandescent filament lamps or glow lamps; the latter are discussed under "Miscellaneous Discharge Lamps" later in this chapter. Also, electroluminescent lamps and light-emitting diodes (see sections below) are included in the miniature lamp family. Incandescent miniature lamps21 and glow lamps22 are specified completely by numbers standardized and issued by ANSI. With the notable exception of multiple holiday lamps, miniature incandescent lamps are designed to operate below 50 V. These voltages are often obtained from batteries or generators. These miniature lamps can be operated on 120-V circuits when step-down transformers, series circuits, rectifiers, or resistors are used to reduce the voltage. Miniature lamps are used chiefly when conditions require a small light source or little power. To insure that the lamp is as small as possible, base sizes are matched to the bulb and the application. They have many uses, principally in automobiles, aircraft, decoration, and as glow lamps in electronic circuits. Both long-life and fleet-service (heavy-duty filaments) products are replacements for standard miniature lamps in the automotive sector. Halogen miniature lamps often are used to replace incandescent miniature lamps. These halogen lamps deliver a whiter and brighter light than the incandescent lamps they replace. Subminiature lamps have increased in popularity. They range in size from T-2 to T-1/8. The T-1-3/4 and T-1 are used extensively for instruments and indicators. The T-5/8 and smaller sizes down to T-1/8 are used chiefly in such novelty applications as tiny flashlights, jewelry, and medical instruments. Train and Locomotive. Lamps designated for train and locomotive service are designed for several classes of low-voltage (75-V or lower) service. The power usually is provided by generators, with a battery connected in parallel so that both supply power to the lamps. Low-voltage lamps have shorter and heavier

filaments than 120-V lamps of the same wattage; consequently, they are more rugged and have higher efficacy. DC Series. Transit system voltages and some railway shop and yard voltages range from 525 to 625 V. Lamps for such service are operated with five to twenty lamps in series. The design voltages of individual lamps operated five in series are nominally 115, 120, and 125 V. Lamps of the dc series are rated in unusual wattages (36, 56, 94, or 101 W). Gas-filled 30-V lamps are used for car lighting. The trolley voltage divided by 30 determines the number of lamps connected in series across the line. These lamps are equipped with short-circuiting cutouts that short-circuit the lamps on burnout, thus preventing arcing and leaving the remainder of the lamps in a given circuit operating. These 30-V lamps are rated in amperes, instead of in the usual watts. Aviation. Lighting for aviation is divided into two classes: lighting on and around airports, and lighting on aircraft. In airport lighting, both multiple and series type lamps are used. Most systems use series lamps of 6.6- and 20-A designs for airport approach, runway, and taxiway lighting. Multiple lamps are used for obstruction, hazard beacon, and airport identification beacon lighting. On aircraft, small and miniature lamps are used for both interior and exterior lighting. Most lamps used in airport lighting are designed to produce a controlled beam of light complying with required standards from such regulators as the U.S. Federal Aviation Authority (FAA) and Transport Canada. Hazard beacons and airport identification beacons, signaling the presence of high obstructions or the whereabouts of the airport, use lamps ranging from 500 to 1200 W. Lamps used on the airport proper range from 10 to 500 W. Lamps used for aircraft lighting are in the miniature classification, except for landing lamps, which can be rated as high as 1000 W. See also "Flashtubes," later in this chapter, for alternative beacon sources. Tungsten-halogen lamps are often provided in place of many regular incandescent types because of their better lumen maintenance and longer life. See Chapter 23, Transportation Lighting, for information on the application of lamps in airports and aircraft lighting. Indicator and Other Service Lamps. Lamps for indicator, radio, and television service usually are operated from low-voltage transformers. Flasher Lamps. Incandescent lamps that flash automatically (Figure 6-25), because of a built in bimetal strip similar to those used in thermostats, are available in several sizes. When the lamp lights, heat from the filament causes the bimetal strip to bend away from the lead-in wire. This breaks the circuit and the lamp extinguishes. As the bimetal strip cools, it returns to its original position against the lead-in wire and lights the lamp. This alternating cooling and heating keeps the lamp flashing. An exception to this is found in a certain type of miniature screw base lamps called the shorting type. The bimetal in this type is so mounted that it shorts across the lead-in wire when hot. If these lamps, which are difficult to distinguish from the opening type, are inserted in sockets intended for the normal flasher lamps, they can drain batteries, blow fuses, overheat wires, or ruin transformers.

Figure 6-25. Typical flasher showing bimetal strip. Typical Uses. Indicator lamps provide a visual indication of existing circuit conditions. They are widely used in fire and police signaling systems, power plant switchboards, production machinery, motor switches, furnaces, and other devices requiring warning or pilot lights. Indicator lamps can be wired with motor or heating elements and can be used to indicate that the current is flowing to an appliance or that it is functioning properly. They often are used in other instrumentation and with photocells and relays. Flashlights, radios, clocks, bicycles, and toys account for more uses. Other applications include the use of miniature lamps for holiday and other festive occasions, and for colorful patio and garden lighting. For garden lighting, low-voltage miniature equipment is available. Flashlight, Handlantern, and Bicycle. These lamps are commonly operated by dry-cell batteries having an open-circuit voltage of 1.5 V per cell for new batteries, dropping to approximately 0.9 V per cell at the end of battery life. This voltage drop results in a reduction in light output in a manner analogous to that shown in Figure 6-19a. Automotive. Lamps for most passenger vehicles, trucks, and coaches operate at 12 V. The power source is a storage battery-rectified alternator system. Head lamps include sealed beam and halogen capsules. SEALED BEAM AND TUNGSTEN-HALOGEN EPOXY SEALED BEAM LAMPS. These lamps contain filaments, lens, and reflectors in a precise, rugged optical package available in a wide variety of sizes and voltages ranging from 6 to 28 V. Sealed beam lamp lenses are made of borosilicate "hard" glass. The reflector is vaporized aluminum on glass and in the incandescent versions is hermetically sealed to the lens cover. The advantages are accurate reflector contour for accurate beam control, precise filament positioning on rugged filament supports, and high efficacy and excellent lumen maintenance. Vaporized aluminum on glass is an excellent reflector, it does not deteriorate, and the normal bulb blackening has little effect on output throughout lamp life. The sealed beam lamp is particularly suitable where a large amount of concentrated light at low voltage is required, as with automotive head lamps. Tungsten-halogen epoxy-sealed beam lamps contain a small bulb of aluminosilicate glass, surrounding the filament. The bulb also contains a high-pressure, rare-gas atmosphere with small additions of halogen compounds as required for operation of the tungsten-halogen cycle. REPLACEABLE AUTOMOTIVE HALOGEN HEAD LAMPS. Changes in automotive styling have led to the creation of replaceable halogen capsules. These lamps meet the current Society for Automotive Engineers (SAE) standards while allowing the designer to use additional shapes of head lamps. Special automotive bases have the advantage that only the light source, rather than the entire assembly, can be replaced when necessary.

Sealed beam lamps are made in a number of rectangular and round versions to allow for automotive design flexibility. They are also available in both single and dual filament designs.

Photo-optical and photographic applications Photo-Optical Lamps. Lamps designed specifically for photo-optical applications typically require tighter tolerances on the positioning of the filament and very often have special prefocus bases to ensure proper alignment in application. In the typical photo-optical application, the prime objective is to collect and direct the lamp's output through an aperture or film gate. In some applications, such as video projection and fiber-optic illuminators, collimation of the light is also needed. In many cases low-voltage filaments are used to provide the smallest but highest luminance source possible, resulting in the greatest optical efficiency. Life is often sacrificed for efficacy and source flex. Typical applications are film, slide, overhead, and video projection, microfilm viewers, and microscope and fiber-optic illuminators. Tungsten-halogen lamps have nearly replaced conventional incandescent lamps for most photo-optical applications. Their small size allows for more efficient optical control. See "Tungsten-Halogen Lamps" earlier in this chapter for additional information. One of the principal developments has been the adoption of halogen lamp types with integrated external dichroic mirrors. By carefully positioning the lamp filament in ellipsoidal or parabolic dichroic reflectors, good beam control is possible. This obviates bulky and expensive external condensers and reflectors. The dichroic mirror is constructed to transmit most of the infrared radiation and reflect light through the film plane. This results in a lower film gate temperature or aperture temperature leading to longer film and optical component life. Photographic. Lamps used specifically for photographic service are adapted to the response or sensitivity of several classes of film emulsions. Some lamps are specified in terms of color temperature, which serves as a basic rating for film exposure data. Lamp life is less important. Lamps of various size often are matched for color temperature, and rated life varies as necessary with wattage to achieve the specified color temperature. Typical color temperature-rated lamps of conventional construction can drop by 100 K throughout life. There is negligible change in the color temperature of tungsten-halogen lamps during their life. Photoflood. These are high-efficacy sources similar to other incandescent filament lamps for picture taking, with color temperatures ranging from 3200 to 3400 K. Because of their high filament temperature, these lamps generally produce approximately twice the luminous flux and three times the photographic effectiveness of similar wattages of general-service lamps. Relatively small bulb sizes are employed. The 250-W No. 1 photoflood, for example, is the same size as the 60-W general-service lamp. These lamps can be conveniently used in less bulky reflecting equipment or for certain effects in ordinary residential or commercial luminaires. The photoflood family includes reflector and projector (PAR) lamps with various beam spreads. Some of these have tungsten-halogen sources; some have integral 5000-K daylight filters. In addition, tungsten-halogen lamps in several sizes and color temperatures are classed as photofloods and used in specially designed reflectors.

FLUORESCENT LAMPS The fluorescent lamp is a low-pressure gas discharge source, in which light is produced predominantly by fluorescent powders activated by UV energy generated by a mercury arc. The lamp, usually in the form of a long tubular bulb with an electrode sealed into each end, contains mercury vapor at low pressure with a small amount of inert gas for starting. The inner walls of the bulb are coated with fluorescent powders commonly called phosphors. When the proper voltage is applied, an arc is produced by current flowing between the electrodes through the mercury vapor. This discharge generates some visible radiation, but mostly invisible UV radiation, the principal lines being approximately 254, 313, 365, 405, 436, 546, and 578 nm. The UV in turn excites the phosphors to emit light. The phosphors are generally selected and blended to respond most efficiently to 254 nm,23,24 the primary wavelength generated in a mercury low-pressure discharge. See Figure 6-26 and the discussion of fluorescence in Chapter 1, Light and Optics. Like most gas discharge lamps, fluorescent lamps must be operated in series with a current-limiting device. This auxiliary, commonly called a ballast, limits the current to the value for which each lamp is designed. It also provides the required starting and operating lamp voltages and may provide dimming control.

Lamp Construction Bulbs. Linear fluorescent lamps are commonly made with straight, tubular bulbs varying in diameter from approximately 6 mm (0.25 in. T-2) to 54 mm. (2.125 in. T-17) and in overall length from a nominal 100 to 2440 mm (4 to 96 in.). The bulb is historically designated by a letter indicating the shape, followed by a number indicating the maximum diameter in eighths of an inch. Hence T-8 indicates a tubular bulb 8/8 in., or 1 in. (26 mm), in diameter. The nominal length of the lamp includes the thickness of the standard lampholders and is the back-to-back dimension of the lampholders with a seated lamp. Fluorescent lamps also come in shapes other than straight tubes. U-shaped tubes are formed by bending tubes in half. They are commonly used in 0.61 m (2 ft) square luminaires. Circular (circline) lamps are tubes bent in a circle with the two ends adjacent to each other. In increasing use are smaller diameter, singleended, compact fluorescent lamps consisting of multiple shaped tubes joined together to form a continuous arc path (Figure 6-26). They are designed to approach the size of the incandescent lamp. Commonly used lamp designations are shown in Figure 6-27. The 32 W T-8 lamp is used as an example, but the designations for most fluorescent lamps follow the same principles. Nomenclature of Fluorescent Lamps. Fluorescent lamps can be designated as illustrated in Figure 6-27. This is only one example; often manufacturers will adopt variations. The specifier should consult with the manufacturer to ensure correct nomenclature for design purposes. Electrodes. Two electrodes are hermetically sealed into the bulb, one at each end. These electrodes are designed for operation as either cold or hot cathodes, more correctly called glow or arc modes of discharge operation. Electrodes for glow (cold cathode) operation can consist of closed-end metal cylinders, generally coated on the inside with an emissive material. Cold cathode lamps operate at a few hundred milliamperes, with a high value of the cathode fall (the voltage required to create ion and electron current flow) in excess of 50 V. The arc mode (hot cathode) electrode generally is constructed from a single tungsten wire, or a tungsten wire around which another very fine tungsten wire has been uniformly wound. The larger tungsten wire is coiled, producing a triple-coil electrode. When the fine wire is absent, the electrode is referred to as a coiledcoil electrode. The coiled-coil or triple-coil tungsten wire is coated with a mixture of alkaline earth oxides to enhance electron emission. During lamp operation, the coil and coating reach temperatures of approximately 1100°C (2012°F), at which point the coil-and-coating combination thermally emits large quantities of electrons at a low cathode fall, in the range of 10 to 12 V. The normal operating current of arc mode lamps is approximately 1.5 A or less. As a consequence of the lower cathode fall associated with the hot cathode, more efficient lamp operation is obtained, and therefore most fluorescent lamps are designed for such operation.

Figure 6-26. Cutaway view of some common fluorescent lamps: (a) A typical rapid-start fluorescent lamp and the production of light; (b) lamp electrode construction; (c) detail of the electrode; (d) a screw-in compact fluorescent lamp with built-in ballast; (e) a 2-pin plug-in compact fluorescent lamp with built-in starter.

Figure 6-27. Nomenclature of Fluorescent Lamps Gas Fill. The operation of the fluorescent lamp depends on the development of a discharge between the two electrodes sealed at the extremities of the lamp bulb. This discharge is developed by ionization of mercury gas contained in the bulb. The mercury gas is typically maintained at a pressure of approximately 1.07 Pa (0.00016 lb/in.2), which is the vapor pressure of liquid mercury at 40°C (104°F), the optimum bulb wall temperature of operation for which most lamps are designed (see the section "Temperature Effect on Operation" below). In addition to the mercury, a rare gas or a combination of gases at low pressure, from 100 to 400 Pascals (0.015 to 0.058 lb/in.2), is added to the lamp to facilitate ignition of the discharge. Standard lamps employ argon gas; energy-saving types, a mixture of krypton and argon; others, a combination of neon and argon or of neon, xenon, and argon. Phosphors. The color of the light produced by a fluorescent lamp depends on the blend of phosphors used to coat the wall of the tube (see Figure 1-12 for a list of important phosphors). Many different white and colored fluorescent lamps are available, each having its own characteristic spectral power distribution (Figure 6-28). These types have a combination of continuous and line spectra.

Figure 6-28a and b. Approximate spectral power distribution charts for various types of fluorescent lamps.

Figure 6-28c. Approximate spectral power distribution charts for various types of HID lamps. Popular fluorescent lamps use three highly efficient narrow-band, rare-earth activated phosphors with emission peaks in the short-, middle-, and longwavelength regions of the visible spectrum. These triphosphor lamps can be obtained with high color rendering, improved lumen maintenance, and good efficacy with correlated color temperatures between 2500 and 6000 K relative to halophosphate lamps.25,26 Since the rare-earth phosphors are expensive, the longer T-5, T-8, T-10, and T-12 triphosphor lamps typically employ a two-coat system consisting of a less expensive halophosphate phosphor applied with the rare-earth type. The rare-earth activated phosphor is closest to the mercury discharge and, as a result, the spectral power distribution of the lamp is more influenced by these phosphors. Common commercial types have correlated color temperatures of 3000, 3500, and 4100 K. A variety of lamp types is available that radiate in particular wavelength regions for specific purposes, such as plant growth, merchandise enhancement, and medical therapy. Various colored lamps, such as blue, green, and gold, are obtained by phosphor selection and filtration through pigments.

Figure 6-29. Typical bases for linear and compact fluorescent lamps (not to scale). ANSI designations are shown. Bases. For satisfactory performance, a fluorescent lamp must be connected to a ballasted electrical circuit with proper voltage and current characteristics for its type. Several fluorescent lamp base designs are used. The bases physically support the lamp in most cases and provide a means of electrical connection (Figure 6-26). Straight tube lamps designed for instant-start operation (see the section "Instant-Start Lamp and Ballast Operation" below) generally have a single connection at each end. As a consequence, a single-pin base is satisfactory. Preheat and rapid-start lamps (see the section "Lamp Starting" below) have four electrical connections, two at each end of the tube, and therefore require dualcontact bases. In the case of the circline lamp, a single four-pin connector is required. Many compact fluorescent lamp bases have unique designs to help ensure their use with the correct ballast. Single-ended compact fluorescent lamps with integral starters have plastic bases containing a glow switch and a noise reduction filter capacitor. These bases have two connection pins. Some lamp wattages are available without the starting components mounted within the bases and have four connection pins (Figure 6-29). Only four-pin lamps are dimmable. For incandescent lamp retrofit applications, self-ballasted compact fluorescent lamps have medium screw bases.

Common Fluorescent Lamp Families T-12 Fluorescent Lamps. Until the National Energy Policy Act of 1992 (EPACT), the most commonly applied fluorescent lamp in the United States and Canada was the T-12, 40-W, 4-ft (1.22-m), rapid-start lamp with a cool white or warm white phosphor. EPACT banned the production of these type lamps after 1995. EPACT also impacted the T-12, 8-ft (2.44 m) lamps. As with 4-ft lamps, only reduced wattage or improved color rendition lamps are currently produced for U.S. consumption. For many new installations, the more efficient T-8 lamps are often specified. Legislation similar to EPACT exists in Canada, where energy efficiency standards for fluorescent lamp ballasts, fluorescent lamps, and incandescent reflector lamps have been established under the Energy Efficiency Act. Regulated products cannot be imported into Canada or traded between its provinces unless they meet the regulatory requirements. Energy-Saving Fluorescent Lamps. In response to the energy crisis of the 1970s, lamp companies introduced halophosphate T-12 lamps filled with an argonkrypton gas mixture, rather than argon only. The 4-ft (1.22 m) lamps can be operated suitably on a ballast designed for 4-ft (1.22 m) 40-W lamps, but because of the different gas mixture they dissipate approximately 34 W per lamp. Any of the energy-saving ballasts that operate standard lamps at full light output can be used, provided the ballast is listed for use with the lamps; this information is stated on the ballast label. These lamps may not be used with any ballast that provides reduced wattage and thus reduced light output in a standard lamp, nor with any ballast that does not list the lamp on its label. A transparent conductive coating is applied to these energy-saving lamps, resulting in a lower required starting voltage and less lumen output. By using these lamps as a retrofit in overilluminated spaces, a saving of 5 to 6 W per lamp can be achieved.

Whether operated on standard or energy-efficient magnetic ballasts, energy-saving fluorescent lamps generate approximately 87% of the light generated by a standard (40-W T-12) lamp at 25°C (77°F). This lamp-ballast system is less efficient than the standard argon gas lamp-ballast system, since it generates fewer lumens per watt. This is due to increased ballast losses. In addition, these lamps cannot be dimmed as easily as standard T-12 lamps, and they are more sensitive to temperature, especially in regard to starting, and should not be started or operated at low temperatures. T-8 Fluorescent Lamps.27,28 T-8 fluorescent lamps are a family of 1-in.-diameter (25.4 mm) straight tube lamps manufactured in some of the same lengths as T-12 lamps. The 4-ft version of the lamp is designed to consume approximately 32 W. It is also available in 2-, 3-, 5- and 8-ft. (0.16-, 0.91-, 1.52-, and 2.44-m) lengths. The smaller diameter makes it economical to use the more efficient and more expensive rare-earth phosphors. Although the T-8 and T-12 lamps are physically interchangeable, they cannot operate on the same ballast. T-8 lamps are designed to operate on line-frequency rapid-start ballasting systems at approximately 265 mA, or on high-frequency electronic ballasts at slightly less current. Due to the higher efficacies that can be reached with T-8 systems, they have replaced the conventional T-12 lamps in many applications. T-5 Fluorescent Lamps. T-5 fluorescent lamps are a family of smaller diameter straight tube lamps employing triphosphor technology. Available only in metric lengths and mini bipin bases, the T-5 lamps provide a higher source brightness than T-8 lamps and better optical control. The lamps provide optimum light output at an ambient temperature of 35°C (95°F) rather than the more typical 25°C (77°F), allowing for the design of more compact luminaires. Also available are high-output versions providing approximately twice the lumens at the same length as the standard versions. T-5 lamps are designed to operate solely on electronic ballasts. Their unique lengths, special lampholder, and ballast requirements make them unsuitable for most retrofit applications. These lamps are used in shallower luminaires than the T-8 lamps, which are more efficient over all than luminaires for T-8 lamps. Compact Fluorescent Lamps. The rare-earth activated phosphor has led to the development of a growing variety of multitube or multibend single-ended lamps known as compact fluorescent lamps (CFLs). The lamps originally were designed to be interchangeable with conventional 25- to 100-W incandescent lamps, but now this lamp type includes sizes that replace conventional fluorescent lamps in smaller luminaires. T-4 and T-5 tubes typically are used in compact fluorescent lamps. There are many techniques of adding, bending, and connecting the tubes to obtain the physical size and lumen output desired. The tube portion of the lamp is sometimes enclosed in a cylindrical or spherical outer translucent jacket made of glass or plastic. Some lamps contain the lamp starter, while others contain both the starter and the ballast, which can take the form of a simple magnetic choke or an electronic ballast. Present compact lamp wattages vary from 5 to 55 W, and rated lumen output ranges from 250 to 4800 lm. Overall lamp length varies from 100 to 570 mm (3.93 to 22.4 in.), depending on lamp wattage and construction. Some designs with self-contained ballasts are equipped with Edison-type screw-in bases for use in incandescent sockets (Figure 6-26d), while other designs use special pin-type bases for dedicated use with mating sockets designed for lamps of a particular wattage (Figure 6-26e). Because of the high power density in these lamps, high-performance phosphors are used extensively in order to enhance brightness, lumen maintenance, and color rendering ability. Amalgams can be added to some versions to enhance performance under a range of operating temperatures. Special Fluorescent Lamps: Subminiature, Reflector, Cold Cathode, and Electrodeless. In addition to the lamps described above, four families of fluorescent lamps are designed and constructed for special applications: subminiature, reflector, cold cathode fluorescent, and electrodeless fluorescent lamps. Subminiature fluorescent lamps are extremely small. They were first used in the backlighting of liquid crystal displays. They are two basic types: hot cathode and cold cathode. The lamps all have bulb diameters of 7 mm (approximately T-2-1/2). The cold cathode series ranges from 1 to 3 W, having an output of 15 to 130 lm, respectively. Standard lamp lengths range from 10 to 50 mm (0.4 to 2.0 in.). These low-power light sources have a low bulb wall temperature, which is important when backlighting displays where space is limited and components must be kept cool. The lamps have a rated life of 20,000 h. The hot cathode family of subminiature fluorescent lamps ranges from 4 to 13 W, with lumen output packages ranging from 95 to 860 lm, respectively. The lumen output is similar to T-5 preheat fluorescent lamps of comparable length. The hot cathode lamps have a rated life of 10,000 h. Their high light output lends them to such general lighting applications as display lighting, valance lighting, furniture-mounted task lighting, and other applications requiring small-diameter, linear light sources. The triphosphor blends in both hot and cold cathode products provide for improved efficacy at the higher wattages and good color rendering (CRI of about 80). Reflector fluorescent lamps are designed for applications requiring directional light output distribution patterns. They have a white powder reflective layer between the phosphor and the bulb that covers a major angular portion of the envelope wall. The major portion of the light is emitted through the strip coated with just the fluorescent phosphor. A cross-sectional diagram and a relative candlepower distribution for a 235° reflector lamp are shown in Figure 6-30. Reflector lamps with other angular widths are available. As a consequence of the reflector layer, absorption of generated light is somewhat higher than in standard fluorescent lamps, producing a somewhat reduced total light output. Cold cathode fluorescent lamps often are used in decorative, sign lighting, and other architectural applications. Due to their high energy losses associated with electrode operation, they are not as efficacious as the more widespread hot cathode lamps for lengths up to 2.44 m (8 ft). The lamps can be custom manufactured in special shapes and sizes. They are frequently manufactured with small diameter tubing so they can be bent into various shapes and sizes. Cold cathode lamps with color phosphors can replace neon tubes in many applications where exposed sources are acceptable. Other advantages of cold cathode lamps include immediate starting, even under cold conditions, and long life unaffected by the number of starts. Neon lamps are cold cathode lamps lacking a phosphor coating. The color of the nonfluorescent neon lamps is determined primarily by the fill gas. Neon emits red, whereas argon mixed with mercury vapor emits blue. Combined with colored glass, these and other gas fills create additional colors. Other special lamps are available for extreme ambient temperatures. One family, designed for low temperatures, incorporates a jacket to conserve heat. Another, for high temperatures, incorporates a mercury amalgam. In both cases, these lamps are designed to optimize the mercury vapor pressures at unusual temperatures. Electrodeless Lamps. Electrodeless lamps have begun to appear on the general lighting market because of advances in the electronics industry and changes in electromagnetic interference (EMI) standards over the last 30 years. Electrodeless lamps use an electromagnetic (EM) field, instead of an electric current passing through electrodes, to excite the gas in a bulb. Electrodeless lamps can be categorized according to the method by which they produce EM fields: either inductive discharge or microwave discharge. Inductive discharge lamps (Figure 6-31), also known as induction lamps, operate using the principle of induction. These lamps also are called electrodeless fluorescent lamps because their EM fields produce light by exciting the same phosphors found in conventional fluorescent lamps. They operate as follows:

Figure 6-30. Cross-section diagrams and relative candlepower distribution curves for 235-degree reflector lamp.

Figure 6-31. Diagram of an induction lamp. 1. The radio frequency (RF) power supply sends an electric current to an induction coil (a wire wrapped around a plastic or metal core). 2. The current passing through the induction coil generates an EM field. 3. The EM field excites the mercury in the gas fill, causing the mercury to emit ultraviolet (UV) energy. 4. The UV energy strikes and excites the phosphor coating on the inside of the glass bulb, producing light. Microwave discharge lamps (Figure 6-32) generate microwaves, which excite the plasma. They operate as follows: 1. A magnetron generates a microwave field. 2. The microwaves travel through a waveguide into a cavity. 3. In the cavity, a hollow glass or quartz ball rotates at high speed to stabilize the fill in the ball (necessary for a uniform light distribution). 4. The excited fill forms a plasma that emits light.

Figure 6-32. Operation of the microwave discharge lamp.

Figure 6-33. Performance Characteristics of Typical Electrodeless Lamps

The efficacies of electrodeless lamps are like those of CFLs or HID lamps of comparable light output. Electrodeless lamps use rare-earth phosphors, giving them color properties similar to those of higher-end fluorescent lamps. Electrodeless lamps are electronic devices, and like all electronic devices they generate EM waves. Electromagnetic interference (EMI) occurs when unwanted EM signals, which can travel through wiring or radiate through the air, interfere with desirable signals from other devices. In the United States, the Federal Communications Commission (FCC) regulates EM emissions in the communication frequencies of 450 kHz to over 960 MHz. Canada also regulates EM emissions over these frequencies through Industry Canada. Manufacturers must comply with FCC regulations to sell products in the United States. However, manufacturer compliance does not assure that EMI will not occur in unregulated frequencies. The International Special Committee on Radio Interference develops standards for EMI from lighting devices, which are accepted by the European community. Of the available induction lamps, one operates at 13.56 MHz and meets FCC requirements without shielding. It is approved for both commercial and residential use. Others operate at 2.65 MHz, which meets the European community's standard for induction lighting. Sometimes, appropriate luminaires must be used to meet shielding requirements; some lamps meet the FCC's EMI requirements for commercial use but not for residential use because their reflectors offer some shielding. Figure 6-33 gives the performance characteristics of typical electrodeless lamps in comparison to other lamps. The microwave lamp operates at 2.45 GHz for regulatory and economic reasons. That frequency is approved for consumer electronics; for example, microwave ovens operate at 2.45 GHz. Because of the popularity of the microwave ovens, magnetron parts are produced in large quantities and are relatively inexpensive.

Performance Parameters Figure 6-3 lists several performance parameters for many common fluorescent lamps. These parameters are discussed in more detail below. Luminous Efficacy: Light Output. Three main energy conversions occur in a fluorescent lamp. Initially, electrical energy is converted into kinetic energy by accelerating charged particles. These in turn yield their energy during particle collision to electromagnetic radiation, particularly UV. This UV energy in turn is converted to visible energy by the lamp phosphor. During each conversion some energy is lost, so that only a small percentage of the input is converted into visible radiation. Figure 6-34 shows the approximate energy distribution in a typical cool white fluorescent lamp.

Figure 6-34. Energy distribution in a typical cool white fluorescent lamp.

Figure 6-35. Efficacy of typical fluorescent lamps as a function of bulb diameter, holding gas fill pressure with arc current constant.

Figure 6-36. Efficacy of a typical halophosphate fluorescent lamp as a function of lamp length. The geometric design and operating conditions of a lamp influence its efficacy. Figures 6-35 and 6-36 show the effect of the bulb design on lamp operation. Figure 6-35 shows that at a constant current, as the lamp diameter increases, the efficacy increases, reaches a maximum, and then decreases.29 The reasons for this phenomenon are two-fold. In lamps of small diameter, an excessive amount of energy is lost by recombination of electrons with ions at the bulb wall. As the bulb diameter is increased, this loss decreases, but losses due to imprisonment of radiation become correspondingly larger.

As shown in Figure 6-36, the length of a lamp also influences its efficacy; the greater the length, the higher the efficacy. This is based on two separate energy losses within the lamp: the energy absorbed by the electrodes, which do not generate any appreciable light, and the energy losses directly associated with the generation of light. The electrode losses are essentially constant, whereas the loss associated with light generation depends on lamp length. As lamp length increases, electrode loss decreases relative to the total losses. The operating voltage of a lamp, like its efficacy, is a function of its length, as shown in Figure 6-37. The characteristic electrode voltage drops for the hot and cold cathode T-8 lamps are indicated by the intersection of the curves with the ordinate corresponding to zero arc length.30 Lamp Life. The lamp life of hot cathode lamps is determined by the rate of loss of the emissive coating on the electrodes (Figures 6-26b and c). Some of the coating is eroded from the filaments each time the lamp is started. Emissive coating also is lost by evaporation during normal lamp operation. Electrodes are designed to minimize both of these effects. The end of lamp life is reached when either the coating is completely removed from one or both electrodes, or the remaining coating becomes nonemissive. Because some of the emissive coating is lost from the electrodes during each start, the frequency of starting hot cathode lamps influences lamp life. The rated average life of fluorescent lamps usually is based on three hours of operation per start (3 h/start). The estimated effect of burning cycles on lamp life normalized to 100% at 3 h/start is presented in Figure 6-38. Cold cathode lamps are not appreciably affected by starting frequency because of the type of electrode used.

Figure 6-37. Operating voltage of typical hot and cold-cathode 26-mm (T-8) fluorescent lamps as a function of arc length.

Figure 6-38. Effect of burning cycles on average lamp life for most popular types of rapid-start fluorescent lamps. All fluorescent lamps follow similar functions depending on the specific lamp and ballast used.

Figure 6-39. Typical mortality curve for a statistically large group of fluorescent lamps (at 3 operating hours per start). Some electronic ballasts have been designed to instant start rapid-start T-8 and T-12 lamps. Typically there is a 25% reduction in lamp life based upon 3 h/start. Many other conditions affect lamp life. Ballast characteristics and starter design are key factors for preheat circuits. Ballasts that neither provide specified starting requirements nor operate lamps at proper voltage levels can greatly affect lamp life. For preheat circuits, starters also must be designed to meet specified characteristics. The electrode heating current in rapid-start lamps is critical and is affected not only by ballasts but also by poor lamp-to-lampholder contact or improper circuit wiring. Improper seating of a lamp in a lampholder can prevent electrode heating. Lamps operating in this mode typically fail within 50 to 500 h. Another factor in lamp life is line voltage. If the line voltage is too high, it can cause instant starting of lamps in preheat and rapid-start circuits. If it is low, slow starting of rapid-start or instant-start lamps, or the recycling of starters in preheat circuits, can result. All of these conditions adversely affect lamp life. A typical mortality curve for a large group of fluorescent lamps is given in Figure 6-39. This curve has recently been validated, on a first order basis, for rapid-cycle switching. Ballasts are available for low-temperature starting of rapid-start lamps. At higher temperatures, lamps operating on these ballasts will start before the electrodes are properly heated, shortening lamp life. Time delay relays are available to ensure proper electrode heating prior to application of ignition voltage to the lamp.

Rh/Rc ratio is correlated with fluorescent lamp life for rapid-start electronic ballasts. Rc is the cold lamp electrode resistance at room temperature (25°C [77°F]). Rh is the hot lamp electrode resistance at the end of the preheat period but before the glow to arc transition. The average electrode temperature before the lamp glow to arc transition (Th) can be calculated using the equation

where Tc is 25°C. This equation is based on the resistance-temperature relationship for tungsten wire. Lamp manufacturers recommend that approximately 700° C is needed to assure minimum sputtering during lamp starting. This electrode temperature correlates to an Rh/Rc ratio of approximately 4.25. For values less than 4.25, sputtering increases and lamp life decreases. In summary, the Rh/Rc ratio appears to be correlated with lamp life based on lamp starting. A low Rh/Rc ratio indicates that the lamp electrodes have not been heated sufficiently during lamp starting, resulting in reduced lamp life. The data shown here support previous recommendations that for rapid-start electronic ballasts, the Rh/Rc ratio should be equal to or higher than 4.25, representing an average electrode temperature of 700°C. Lumen Depreciation. The light output of fluorescent lamps decreases with accumulated operating time because of photochemical degradation of the phosphor coating and glass tube and the accumulation of light-absorbing deposits within the lamp. The rate of phosphor degradation increases with arc power and decreases with increased coating density. Lamp lumen depreciation (LLD) curves for different fluorescent lamps are shown in Figure 6-40.31,32 Protective coatings are sometimes used to reduce the phosphor degradation. Triphosphors are more stable and allow higher loading levels, as for example in T-5, T-8, compact, and subminiature lamps. The deposit of electrode coating material evaporated during lamp operation causes end darkening. This reduces UV radiation into the phosphors, thereby reducing light output near the ends.

Figure 6-40. Fluorescent lamp lumen depreciation (based on operating on suitable ballast at 3 hr per start). Spectral Power Distribution and Chromaticity. Spectral power distribution data for several fluorescent lamps are shown in Figure 6-28. For a discussion of color and color rendering index, see Chapter 4, Color. Temperature Effect on Operation. The luminous performance, light output, and color of a fluorescent lamp are dependent on the mercury vapor pressure within the lamp, which depends on temperature (Figure 6-41). A fluorescent lamp contains a larger quantity of liquid mercury than will become vaporized at any one time. The excess liquid mercury condenses at the coolest point or points on the lamp. The mercury pressure within the lamp depends on the temperature of the coldest point or points. Lamp construction, design, and wattage, as well as luminaire design, ambient temperature, and wind or draft conditions, affect the cold point. Lamps using mercury amalgams are available for extending the usable ambient temperatures to higher values. An amalgam is an alloy of mercury and other metals. The amalgam stabilizes and controls the mercury pressure. Placed in a fluorescent lamp, it determines the mercury vapor pressure in the discharge by absorbing or releasing mercury. Amalgams typically are used with compact fluorescent lamps where the bulb wall gets so hot that conventional temperature control techniques are less effective. An amalgam keeps mercury pressure in the discharge close to its optimal value as the lamp temperature changes. As a result, an amalgam lamp can produce more than 90 percent of its maximum light output over a wide temperature range. Amalgam lamps also tend to maintain relatively constant light output at different operating positions compared to non-amalgam lamps. However, amalgam lamps can take longer to reach their full light output when turned on.33 The internal temperature of a luminaire can adversely affect the life of some types of fluorescent lamps. High ambient temperatures not only lower the lamp's lumen output but also can change the lamp's electrical characteristics, bring these characteristics outside the design range of the ballast, and therefore allow more than rated current to flow. Long-term operation at higher currents shortens the life of the lamp. As the temperature of the cold point changes, both the light output and the active power also change. Both active power and light output have optimum temperatures. Lamp efficacy, defined as light output divided by active power, is typically maximized at approximately 40°C (104°F) (Figure 6-41). Since temperatures within luminaires are typically above the optimum temperature for the lamps and since light loss beyond the optimum temperature is nearly linear, a rule of thumb can be used to estimate light loss as a function of high ambient temperatures. There will be a 1% loss in light output for every 1.1°C (2°F) increase in the ambient temperature above 38°C (100°F).33

Figure 6-41. Typical fluorescent lamp temperature characteristics. Exact shape of curves will depend on lamp and ballast type; however, all fluorescent lamps have curves of the same general shape, since this depends on mercury vapor pressure.

Figure 6-42. The light output characteristics for a nonamalgam compact fluorescent lamp show that the cool zone designed into the lamp geometry to help lower the minimum bulb wall temperature is most effective when the lamp operates base-up.

Figure 6-43. Comparison of relative light output vs. ambient temperature for two compact fluorescent lamp designs: one with amalgam (curve a) and one nonamalgam (curve b) in a base-up burning position. Compact fluorescent lamps often are more sensitive to the temperature of the operating environment than the standard straight tube lamps. In some cases the increased temperature inside the luminaire results in lower than rated light output. Additionally, the performance of many types of compact fluorescent lamps depends on their operating position. Figure 6-42 shows the influence of operating position on typical compact lamps. Some compact fluorescent lamps employ amalgam technology that reduces the lamp sensitivity to burning position and lumen loss due to high and low temperature (Figure 6-43). Most T-8 and T-12 lamps, which are intended primarily for indoor use, have been designed for their light output and luminous efficacy to reach optimum values at a minimum bulb wall temperature of 38°C (100°F). In well-designed luminaires, this temperature is typically reached when the lamps are operated at rated power under usual indoor temperatures. Curves for an 800-mA high-output fluorescent lamp are shown in Figure 6-44 (left). As these curves indicate, the light output falls to very low values at temperatures below freezing. Lamps intended for indoor operation display poor low-temperature performance unless protected by suitable enclosures. Figure 644 (right) shows the relationship between ambient temperature and light output for a typical outdoor floodlight using 800-mA high-output lamps. While considerable variation occurs with temperature change, satisfactory illumination is obtained for most winter temperatures. Each lamp-luminaire combination has its own distinctive characteristic of light output as a function of ambient temperature. In general, the shape of the curve is quite similar for all luminaires, but the temperature at which the highest light output is reached can be different. Effects of Temperature on Color. The color of light from a fluorescent lamp depends on the phosphor coating and also the mercury arc discharge. Each of these components reacts differently to temperature changes. Figure 6-45 shows a typical color shift characteristic of a halophosphate fluorescent lamp.

Figure 6-44. Light output versus ambient temperature. (Left) F96T12/HO fluorescent lamp. Light output falls to low values at temperatures below freezing. Loss in light at high ambient temperatures is much less. (Right) Two F72T12/HO lamps mounted in a typical floodlight. Performance of lamps designed for indoor application is considerably improved when operated in a suitable enclosure.

Figure 6-45. This CIE chromaticity diagram contains data for four different halophosphate fluorescent lamps. It shows that the color each lamp produces shifts toward the blue/green with increasing temperature. The lowest point on each curve is at −20°C ( −4°F). Following each curve up and to the left as it bends over, the furthest point is at 120°C (248°F). Intermediate points are 20°C (36°F) apart. The chromaticity coordinates of any point are obtained from the x and y axis. Color shift can be a concern when substantial differences in internal temperature exist between adjacent luminaires. This can arise from the proximity of certain luminaires to air diffusers or open windows; differences in ceiling cavity conditions or ceiling material with surface and recessed equipment; differences in the tightness of enclosures with enclosed equipment; differences in lamp loading or number of lamps in identical luminaires; and use of some of the luminaires as air diffusers in the air-conditioning system. High-Frequency Operation of Fluorescent Lamps. High-frequency electronic ballasts generally provide power to the fluorescent lamp in the range of 10 to 50 kHz from a 50- to 60-Hz power supply. The primary advantage of high-frequency electronic ballasts for fluorescent lighting systems is higher efficacy relative to the 60-Hz magnetic ballast systems. As shown in Figure 6-46, efficacy increases rapidly with high-frequency operation until 20 kHz; in the range from 20 to 100 kHz, efficacy is constant. The improved performance of the fluorescent lamp at high frequencies has been attributed to two factors. First, a reduction in end losses is achieved by elimination of the oscillation on the anode half of the operating cycle. Second, an increase in efficiency of the lamp's positive column (major portion of the arc stream) is achieved by operating at lower wattage. In order to save energy, fluorescent lamps normally are operated at lower than rated wattage with high-frequency electronic ballasts while maintaining the lamp's rated lumen output. In order to avoid audible noise, most electronic ballasts operate the lamp above 20 to 30 kHz. Another consideration in high-frequency operation of fluorescent lamps is radiated and conducted radio-frequency (RF) noise. The electronic ballast must have filter circuitry to constrain the conducted RF to within government regulations. In addition, the lamp current waveform must be chosen to limit RF intensity.

Figure 6-46. Lamp efficacy gain at constant lumen output vs. operating frequency for a 40-watt, T-12 rapid-start lamp.

Figure 6-47. Typical radio interference filter.

Radio Interference. The mercury arc in a fluorescent lamp emits electromagnetic radiation. This radiation can be received by nearby radios, causing an audible noise. Radio noise reaches the receiver either by radiation to the antenna or by conduction over the power lines. Because of the frequencies generated by the fluorescent lamp, radiated interference is ordinarily limited to the AM broadcast band and nearby amateur and communications bands. Radiated interference can be eliminated by moving the antenna farther from the lamp. A distance of 3 m (9.8 ft) is usually sufficient. Where this is not practical, shielding media, such as electrically conducting glass or certain louver materials, can suppress the noise below the interference level. FM, television, and higher frequencies rarely are affected by radiated interference but can be affected by conducted interference. Conducted interference can be suppressed by an electric filter in the line at the luminaire. Figure 6-47 shows a typical design. Luminaires with this type of filtering and appropriate shielding material have been qualified under pertinent military specifications for sensitive areas. Most instant-start ballasts and starters for preheat circuits have capacitors for reducing radio interference. Infrared Interference with Compact Fluorescent Lamps. The use of infrared (IR) radiation for transmitting data and control signals has increased in popularity for equipment such as television and video cassette recorders, computers, and medical devices. Such equipment generally uses IR receivers that, in North America, operate at carrier frequencies from 33 kHz to 40 kHz, and some at 56 kHz. With the increased use of compact fluorescent lamps operated on electronic ballasts, more interference problems have been reported. If the operating frequency of the compact fluorescent ballast or its second harmonic, which is the power supply frequency, is within the frequency band of the appliance IR receiver, interference can occur. To reduce or eliminate this interference, one should move the lamp to a new position or use other lamp and ballast combinations where frequency matching does not occur. Current electronic ballast designs take into account this potential interference by using operating frequencies that minimize this interaction. Additionally, IR receivers now contain improved coding that is less sensitive to stray IR radiation. Flicker and Stroboscopic Effect. The light output of a fluorescent lamp varies with instantaneous power input. Operating on a magnetic ballast with a 60-Hz power input frequency, the resulting 120-Hz variation coupled with phosphor persistence makes the fluctuating light output too rapid for most people to perceive. This assumes, however, that the power input is free of electrical noise. The presence of electrical noise from other equipment can result in frequencies that manifest themselves as visible flicker. Under noise-free operating conditions, the flicker index for typical fluorescent lamps operated with electromagnetic ballasts ranges from 0.01 to approximately 0.1, and is much lower when operated with high frequency electronic ballasts. For a discussion about flicker and the stroboscopic effect, including the definition of the flicker index, see Chapter 2, Measurement of Light and Other Radiant Energy.

Lamp Operation and Auxiliary Equipment General. Like most arc discharge lamps, fluorescent lamps have a negative volt-ampere characteristic and therefore require an auxiliary device to limit current flow. This device, called a ballast, might also provide a voltage sufficient to start the arc discharge. This voltage can vary between 1.5 to 4 times the normal lamp operating voltage. The life and light output ratings of fluorescent lamps are based on their use with ballasts providing proper operating characteristics, which have been established in the ANSI standards for dimensional and electrical characteristics of fluorescent lamps (C78 Series). Ballasts that do not provide proper electrical values might reduce either lamp life or light output or both. This auxiliary equipment requires electrical power and therefore reduces the system efficacy below that based on the power requirements of the lamp. Lamp Starting. The starting of a fluorescent lamp occurs in two stages. First, the electrodes must be heated to their emission temperatures. Second, a sufficient voltage must exist across the lamp to ionize the gas in the lamp and develop the arc. In some starting systems, a voltage is applied between one of the electrodes and ground to help ionization. As the ambient temperature is reduced, it becomes more difficult to start fluorescent lamps. Higher voltages are required to reliably start lamps at low temperatures. For efficient lamp and ballast operation, specific ballasts are generally available for each of the following temperature ranges: above 10°C (50°F) for indoor applications, above − 18°C (0°F) for outdoor temperature applications, and above − 29°C (− 20°F) for outdoor temperature applications. Three different means of starting lamps with magnetic ballasts have been developed. Preheat starting requires an automatic or manual starting switch. Instant starting requires a high ballast open circuit voltage. Rapid starting, the most commonly used starting circuit, continuously heats the electrodes, obviating high voltages and starting switches. Several magnetic circuits are shown in Figure 6-48. In general, for operation on magnetic ballasts, there are differences between lamp designs for different starting methods; therefore, it is important to match the lamp to the starting circuit. The lamp description normally identifies the proper circuit, that is, preheat, rapid, or instant start. For electronic ballasts, new techniques for starting fluorescent lamps have been developed. Electronic ballasts have been introduced that can instant start most rapid-start fluorescent lamps. Additionally, hybrid electronic starting methods are available that combine characteristics of rapid and preheat starting. Electronic ballasts are also available with a soft starting sequence, which is designed to minimize damage to the electrodes during starting and therefore to lengthen lamp life.

Figure 6-48. Typical fluorescent lamp circuits. Preheat Lamp and Ballast Operation. In preheat circuits, the lamp electrodes are heated before application of the high voltage across the lamp. Lamps designed for such operation have bipin bases to facilitate electrode heating. Many preheat-starting compact fluorescent lamps have the starting devices built into the lamp base. The preheating requires a few seconds, and the necessary delay usually is accomplished by an automatic switch that places the lamp electrodes in series across the output of the ballast. Current flows through both electrode filaments, heating them. Subsequently, the switch opens, applying the voltage across the lamp. Due to the opening of the switch under load, a transient voltage (an inductive spike) is developed in the circuit, which aids in ignition of the lamp. If the lamp does not ignite, the switch closes and reheats the filaments. In some systems, preheating is accomplished by a manual switch. The automatic switch is commonly called a starter. It can incorporate a small capacitor (0.006 µF) across the switch contacts to shunt high-frequency oscillations that might cause radio interference. Ballasts are available to operate some preheat lamps without the use of starters. These ballasts use the rapid-start principle of lamp starting and operating and popularly are called trigger start ballasts. Starters For Preheat Circuits. The operation of a preheat circuit requires heating of the electrodes prior to application of voltage across the lamp. Preheating can be effected by use of a manual switch or a switch that is activated by application of voltage to the ballast circuit. A number of automatic switch designs are commercially available. Diagrams for two designs are presented in Figure 6-49. Thermal Switch Starter. A diagram of a thermal switch starter is presented in Figure 6-49a. Initially the silver-carbon contact of the thermal starter is closed, placing the electrodes in series with the parallel combination of the bimetal and the carbon resistor. Upon closing the ballast supply circuit, the output voltage of the ballast is applied to this series-parallel wiring combination. The current heats the bimetallic strip in the starter, causing it to open the silver-carbon contact. The time of opening is sufficient to raise the temperature of the electrodes to approximately its normal operating value. Upon opening the circuit, the ballast output voltage in series with an inductive spike (kick) voltage is applied to the lamp. If the lamp ignites, its normal operating voltage maintains a low current through the carbon resistor, developing and transferring sufficient heat to the bimetal to hold its contact open thereafter.

Figure 6-49. Starter switches for preheat cathode circuits: (a) thermal type; (b) glow switch type. Should the lamp fail to start on the first attempt, the ballast open-circuit voltage applied to the carbon resistor heats the bimetal sufficiently to cause the silver contact to move against the third contact. This short-circuits the carbon resistor, permitting preheating current to flow through the electrodes. As the bimetal cools, the circuit through the third contact is opened, resulting in the application of the circuit voltage to the lamp again. This making and breaking of the circuit through the third contact continues until the lamp ignites. The bimetal circuit is held open thereafter as noted above. The carbon contact circuit functions only when the line voltage is initially applied to the ballast. Thermal-switch starters require some power (0.5 to 1.5 W) during lamp operation, but their design ensures positive starting by providing an adequate preheating

period, a high induced starting voltage, and characteristics inherently less susceptible to line voltage variations. For these reasons they give good all-around performance under adverse conditions, such as direct-current operation, low ambient temperature, and varying voltage. Glow-Switch Starter. The circuit for this starter is presented in Figure 6-49b. The bulb is filled with an inert gas chosen for the voltage characteristics desired. On starting, the line switch is closed. There is almost no voltage drop in the ballast, and the voltage at the starter is sufficient to produce a glow discharge between the contacts. The heat from the glow distorts the bimetallic strip, the contacts close, and electrode preheating begins. This short-circuits the glow discharge so that the bimetal cools, and in a short time the contacts open. The open-circuit voltage in series with an inductive spike voltage is applied to the lamp. If the lamp fails to ignite, the ballast open-circuit voltage again develops a glow in the bulb, and the sequence is repeated until the lamp ignites. During normal operation, there is not enough voltage across the lamp to produce further starter glow, so the contacts remain open and the starter requires no power. Cutout Starter. This starter resets either manually or automatically. It is designed to prevent repeated blinking or attempts to start a deactivated lamp. This type of starter should be good for at least ten or more renewals. Lamp Failure in Preheat Circuit. Starters that provide no means for deactivation when a lamp fails will continue to attempt to start the lamp. The lamp might repeatedly blink on and off, and the ballast or starter will eventually fail. Thus it is important to remove a failed preheat lamp immediately. Instant-Start Lamp and Ballast Operation. Arc initiation in instant-start lamps depends solely on the application of a high voltage across the lamp. This voltage (400 to 1000 V) ejects electrons from the electrodes by field emission. These electrons flow through the tube, ionizing the gas and initiating an arc discharge. Thereafter, the arc current provides electrode heating. Instant-start lamps need only a single contact at each end. A single pin is used on most instant-start lamps. These are commonly called slimline lamps. A few instant-start lamps use bipin bases with the pins connected internally. In the case of lamps designed for instant starting at 400 to 1000 V open circuit, it is necessary to provide some means of counteracting the effect of humidity on the capacitive lamp-ground current that initiates the necessary glow discharge. Most manufacturers coat the outside of bulbs of this type of lamp with a transparent, nonwetting material; others apply a narrow conducting strip along the bulb. A grounded conducting plate, such as a metal reflector near the lamp, commonly known as a starting aid, is necessary to obtain the lowest lamp starting voltage.34 Rapid-Start Lamp and Ballast Operation.35-38 Lamps designed for rapid-start operation typically have low-resistance cathodes. Normally, the cathodes are heated continuously by the application of cathode voltage while the lamps are in operation. In some energy-saving circuits, the cathode voltage is reduced or disconnected after the starting of the lamps. Heating is accomplished through low-voltage windings built into the ballast or through separate low-voltage transformers designed for this purpose. This results in a starting-voltage requirement similar to that of preheat lamps. Lamps usually start in approximately one second, which is the time required to heat the filaments to their proper temperature. A starting aid, consisting of a grounded conducting plate, is required for reliable starting. For lamps operating at 500 mA or less, the nominal distance between the lamp and a 25 mm (1-in.) wide conducting plate is 13 mm (0.5 in.); for lamps operating at currents greater than 500 mA, the nominal distance to the conducting strip is 25 mm (1 in.). Rapid-start lamps are coated with a transparent nonwetting material to counteract the adverse effect of humidity in lamp starting. All 800-mA and most 1500mA lamps operate on the rapid-start principle. Forty-watt and circline lamps designed for rapid-start service also can be used in comparable preheat circuits. Electronic Ballast Starting. All three starting methods are employed in electronic ballasts. When applied to electronic ballasts, the differences between the rapid-start and the preheat technique become less significant. The preheat designs use internal timing components that delay the full open-circuit voltage while applying power to the electrodes for preheating. After starting, the power to the electrodes is reduced to almost zero. Most of the rapid-starting systems do not rely on a potential to ground to aid the start as with the magnetic ballast. The starting method is similar to the preheat technique except that the electrode voltage usually remains after the lamp starts. Electronic ballasts have been designed to "instant-start" rapid-start fluorescent lamps. Typically there may be a reduction in expected life when operated in this manner. Ballasts Magnetic Ballasts. The construction of a typical thermally protected rapid-start magnetic ballast is shown in Figure 6-50. The components include a transformer-type core and coil. A capacitor might be included. These components are the heart of the ballast, providing sufficient voltage for lamp ignition and lamp current regulation through their reactance. The core-and-coil assembly is made of laminated transformer steel wound with copper or aluminum magnet wire. The assembly is impregnated with a nonelectrical insulation to aid in heat dissipation and, with leads attached, is placed in a case. The case is filled with a potting material (e.g., hot asphalt) containing a filler such as silica. This compound completely fills the case, encapsulating the core and coil and the capacitor. The base is then attached.

Figure 6-50. Components of a magnetic ballast (a) and high-frequency electronic ballast (b) for a linear fluorescent lamp.

The average ballast life at a 50% duty cycle and a proper ballast operating temperature is normally estimated at twelve years. Ballast life is reduced by higher temperatures or longer duty cycles. Most fluorescent lamp ballasts used indoors should have an internal thermal protection device. This prevents misapplication of the ballast in high-temperature applications and protects against failure and undesirable conditions that can occur at end of ballast life. In the United States the thermally protected ballast approved by Underwriters Laboratories is known and marked or labeled as "Class P." Because of the magnetic components in most ballasts, including some electronic ballasts, vibrations can cause audible noise. The noise depends on several factors, including the construction, mounting, number, and spacing of the ballasts and luminaires, and the acoustical characteristics and use of the room. Ballast manufacturers publish sound ratings that indicate the relative sound-producing potential of their different models. These ratings are based on the experience of the model in different ambient conditions, operating in places ranging from a quiet school room, church, or office to a relatively noisy business office, store, or factory. An A rating means the ballast/luminaire hum probably will not be noticeable in quiet spaces. A B rating is borderline in the quiet applications, but will not be a problem in the noisier places. The C rating should be used only in noisy places like factories. Electronic Ballasts.39 The operating frequency of electronic ballasts is chosen to be high enough to increase the lamp efficacy and to make ballast noise inaudible, but not so high as to cause electromagnetic interference (EMI). Electronic ballasts also provide a level of light output regulation that is unavailable in totally passive, magnetic ballasts. Designs are available for the rapid- and instant-start of lamps. Some electronic ballasts are designed to operate up to four lamps each. Many are made in the same size and shape as magnetic units in order to ease direct replacement. Some designs have circuits that keep the linecurrent harmonic distortion below 20% and provide a power factor in excess of 90%. Some ballasts also employ circuits to limit the current in-rush when power is applied to the ballast. This in-rush, which is a result of the electrolytic capacitors of an electronic ballast charging up, occasionally had been reported as a problem for switches and relays used to turn power to the ballasts on and off. Electronic ballasts also can be designed to operate with dc and low-voltage systems for applications in buses, airplanes, trailers, and battery-operated emergency systems. In addition, it has been reported that some T-5 and smaller fluorescent lamps have problems at the end of life in field applications. When a fluorescent lamp fails, typically one of the electrodes becomes an open circuit. This can create an asymmetric lamp current operating condition that can produce high local heating, which in turn can crack the lamp bulb or overheat and deform the lamp base. ANSI specifications for magnetic ballasts40 have been revised to address the end-of-life operation of these lamps. Reduced-Wattage Ballasts. Ballasts are available that operate standard lamps at 50 to 80% of their rated wattage. Energy-saving lamps should not be used in combination with these ballasts, since the arc will tend to waver. Energy-Saving Ballasts. Energy-saving ballasts have lower power losses than the more common magnetic ballasts. These may be rated by Certified Ballast Manufacturers (CBM) and are used either with common lamps or with reduced-wattage lamps. For example, power losses in two-lamp 40-W rapid-start ballasts have been reduced by 4 to 5 W per lamp over common magnetic ballasts. A typical two-lamp 40-W unit with a low-loss energy-saving ballast dissipates approximately 86 W, compared to approximately 95 W for most magnetic ballasts. Energy-Saving Systems. Specialized lamp-and-ballast combinations are available to achieve energy savings. These include a 32-W T-8 (4-ft) lamp with a highefficiency ballast and a 28-W T-12 lamp, also with a high-efficiency ballast, having internal solid-state switches that turn off the usual rapid-start cathode heater voltage. These ballasts can also operate a 34-W reduced-wattage lamp. Power reducers are also available for saving energy. These solid-state electronic devices are wired in series with the lamp ballast to reduce operating wattage. Note that a reduction in light output results.41 Ballast Power Factor. The power factor is defined as the ratio of input wattage to the product of root mean square (rms) voltage and rms current. It represents the amount of current and voltage that the customer is actually using as a fraction of what the utility must supply. High power factor is defined as being above 90%. A ballast with low power factor draws more current from the power supply and therefore, larger supply conductors might be necessary. Low-power-factor ballasts are more common with compact fluorescent systems than for 4-ft and 8-ft fluorescent systems. Some public utilities have established penalty clauses in their rate schedule for installations with low power factor. Some utilities require high power factor equipment. Ballast Factor and Ballast Efficacy Factor. The ballast factor is defined by ANSI (ANSI C82.2-1984)40 as the relative light output of a lamp operated on the ballast with respect to the same lamp on a reference ballast, usually expressed in percent. The reference ballasts are discussed in detail for each fluorescent lamp type in the applicable ANSI lamp standards. The ballast efficacy factor (BEF) is defined as the ballast factor in percent, divided by the total input power in watts. In the United States, federal regulation sets limits on the BEF of some ballasts for 1.22 m (4-ft) and 2.44 m (8-ft.) fluorescent lamps, summarized in the following table:

Figure 6-51. General Characteristics of Commonly Used Ballasts

Figure 6-52. Fluorescent Lamp Electrical Characteristics In addition to the federal regulation, some states might impose additional restrictions on the above or on additional lamp types. Specifically excluded are dimming ballasts, ballasts intended for use in ambient temperatures of − 17.8°C (0°F) or lower and ballasts with power factor less than 90% that are designed for residential use (in buildings up to three stories). Moreover, some utility companies have specified a minimum BEF and ballast factor in their rebate programs for energy-efficient equipment. They generally pertain to the same lamps as the U.S. federal regulation. Harmonics. Line-current harmonics are those components of the line current that oscillate at low integer multiples of the fundamental frequency of the power supply. For instance, in North America, the fundamental frequency is 60 Hz, the second harmonic is 120 Hz, the third harmonic is 180 Hz, and so forth. Switching in modern solid-state electronic ballasts can cause substantial line-current harmonics when corrections are not implemented in the ballast. This can be especially harmful in three-phase installations if the third-harmonic current is large. The third harmonic and its multiples add in the neutral wire, while the fundamental currents tend to cancel one another there. If the third harmonic is 33.3% of the fundamental, then the total third harmonic in the neutral wire is equal to the fundamental in the phase wires. This can cause problems, including overheating, if the neutral wire is not properly sized. For these reasons, ANSI C82.1142 places limits on the harmonic content in the line current for electronic ballasts employed in commercial and residential lighting applications. In the United States, several utility companies have included harmonics as an issue in their rebate programs intended for customers who purchase energy-efficient equipment. Early electronic ballasts typically had high total harmonic distortion (THD) content, which led to the development of these standards. Figures 6-51 and 6-52 list THD and electrical characteristics of fluorescent lamps, respectively. Fluorescent Lampholders. Lampholders are designed for each lamp base style. Typically several versions of each are available to allow various spacings and mounting methods in luminaires (Figure 6-53). Proper spacing should be maintained between lampholders in luminaires to ensure satisfactory electrical contact. Manufacturers' catalogs should be consulted for dimension and spacing information on any particular lampholder type. When fluorescent lamps are used in circuits providing an open-circuit voltage in excess of 300 V, or in circuits that permit a lamp to ionize and conduct current with only one end inserted in the lampholder, electrical codes usually require some automatic means for opening the circuit when the lamp is removed. This usually is accomplished by the lampholder so that on lamp removal, the ballast primary circuit is opened. The recessed contact bases on 800- and 1500-mA fluorescent lamps eliminates the need for this disconnect feature in lampholders for these lamps.

Figure 6-53. Typical lampholder designs. Lamp bases for many compact fluorescent lamps are constructed with unique pin-and-keyway systems to prevent installing the wrong lamp. Universal lampholders that allow lamps of any wattage to fit should be avoided. Dimming of Fluorescent Lamps.43-45 Many types of fluorescent lamps are suitable for dimming. Dimming fluorescent lamps differs from dimming incandescent lamps in two key ways. First, fluorescent dimmers do not provide dimming to zero light as do incandescent dimmers. However, products are available to dim to as low as 0.5 to 25% of maximum light output. Second, when dimming fluorescent lamps, the correlated color temperature varies

substantially less over the dimming range than incandescent lamps. Dimming is achieved by reducing the effective lamp current. When doing so, it is necessary to supply the full starting voltage and to maintain the restriking voltage necessary at each 60-Hz half cycle. This is especially true when operating the lamp at low light output. It is also necessary to provide filament heating for all except cold cathode lamps in order to maintain the required electron emissions from the electrodes at all intensities. Early magnetic dimming ballasts achieved dimming by lowering the primary voltage to the ballast transformer. Such a dimming system can be used with twopin cold cathode fluorescent lamps in a series circuit as shown in Figure 6-54. With this arrangement, it is possible to reduce the luminous intensity to approximately 10% of maximum light output. The performance of magnetic dimming ballasts can be improved by adding one or more of the following features: filament transformers to provide filament heat, pulse networks to provide the required restriking voltage at low intensities, and high-frequency keep-alive current to maintain the discharge when the line current is interrupted during part of the 60-Hz cycle. These components can be packaged together with the magnetic ballast or be separately installed in the luminaires as dimming adapters.

Figure 6-54. Typical dimming circuit for series-connected cold-cathode lamps. Controls for the magnetic dimming ballasts are available in two-wire and three-wire configurations. The two-wire systems have a limited dimming range, typically 25% light output at the low end, but can be retrofitted with existing wiring in most cases. The three-wire configurations bring a control wire to an external control, such as a wall station. This flexibility allows for a substantial improvement in the dimming performance. For electronic dimming ballasts, there are several control schemes available. Some manufacturers use a three-wire scheme as described above, so that electronic dimming ballasts can directly replace magnetic dimming ballasts without affecting the control unit. Other three-wire implementations also require the control to be changed. Yet other manufacturers use four-wire systems where two of the wires are used for the dimming signal and the other two to carry the main lamp current. Most currently available fluorescent lamp dimming systems incorporate electronic ballasts that use high-frequency (typically 20 to 50 kHz) switching of the lamp current. They are designed to be used with four-pin, rapid-start, and compact fluorescent lamps and are available for several lamp diameters and lengths. Electronic dimming ballasts generally are more efficient and less bulky than their autotransformer predecessors. Furthermore, lamp flicker can be substantially reduced with electronic dimming ballasts. Most electronic ballasts offer energy savings approximately proportional to the reduction in light output (Figure 6-55). This is particularly true at dimmer settings above 25 to 50% luminous output. Furthermore, four-pin construction allows cathode heating when dimming. This is important since it extends lamp life and eliminates flicker when properly implemented. It is also advisable to select premium-quality knife-edge sockets rather than leaf-spring contacts. This ensures that cathode heating is reliably supplied. Finally, solid-state dimmers are substantially quieter (less humming) than their magnetic predecessors.

Figure 6-55. Light output vs. input power for a typical 40-watt 120-volt rapid-start fluorescent dimming system. The performance of a fluorescent dimming system might not be satisfactory if the lamp is not correctly matched with the dimming ballast and the controller. In particular, reduced wattage, energy-saving retrofit lamps should not be used in dimming systems, unless so recommended by the dimmer manufacturer. Doing so might shorten the life of the lamp and ballast. Flashing of Fluorescent Lamps.46,47 Cold cathode and rapid-start or preheat-start hot cathode fluorescent lamps can be flashed and still maintain good performance. Cold cathode lamps are flashed through control of either the transformer primary or secondary voltage. Hot cathode lamps can be flashed by means of a special ballast that turns the arc current on and off but keeps the cathode heating on. An external flashing device is required with either system. This unit must be rated for the voltage and current involved, and it is recommended that separate contacts be used for each ballast to prevent circulating currents between ballasts. Flashing of fluorescent lamps is sometimes used in advertising.

HIGH-INTENSITY DISCHARGE LAMPS High-intensity discharge (HID) lamps include the groups of lamps commonly known as mercury, metal halide, and high-pressure sodium. The light-producing element of these lamp types is a wall-stabilized arc discharge contained within a refractory envelope (arc tube) with wall loading in excess of 3 W/cm2 (19.4 W/in.2).

Lamp Construction and Operation All high-intensity discharge lamps produce light by means of an electrical arc discharge contained in an arc tube inside the bulb. The arc tube contains tungsten electrodes that terminate the arc discharge at each end of the arc tube. The arc tube also contains a starting gas that is relatively easy to ionize at low pressure at normal ambient temperatures. This starting gas is usually argon or xenon or a mixture of argon, neon, or xenon, depending on the type of HID lamp. The arc

tube also contains metals or halide compounds of metals that, when evaporated into the arc discharge, produce characteristic lines of radiant energy. Each type of HID lamp produces light related to the type of metal that is contained in the arc. Mercury vapor lamps produce light by exciting mercury atoms; highpressure sodium lamps produce light by exciting sodium atoms; and metal halide lamps produce light by exciting several different atoms and molecules, primarily sodium, scandium, thulium, holmium, and dysprosium. The arc tube is contained inside a soft or hard glass outer bulb to protect the arc tube and internal electrical connections from the environment. The outer bulb absorbs the majority of UV energy radiated by the arc tube while allowing light to pass through. The outer glass bulb can be coated with a diffusing material to reduce the source brightness of the lamp. In mercury vapor and metal halide lamps, this diffusing coating can be a color-correcting phosphor that uses UV energy radiated by the arc tube to improve the lamp's overall color rendering properties. Within the outer bulb there are wires suitable for high temperatures to conduct electricity to the arc tube and structural components to support the arc tube. There might be other components, including resistors or diodes used to help start the arc discharge, and devices called getters to purify the atmosphere in the outer lamp. The atmosphere in the outer bulb might be a low-pressure gas (usually nitrogen) or, in many cases, a vacuum. HID lamps have screw bases (medium or mogul) made from brass, nickel, or special alloys to minimize corrosion. Some HID lamps have special bipin bases or pairs of single contact bases at each end of the lamp to provide electrical connections (Figure 6-56). If the outer bulb is broken and the arc tube continues to operate, the lamp emits a significant amount of UV energy. Exposure to people beyond about 15 minutes can produce severe erythemal effects (skin reddening) or eye damage (see Chapter 5, Nonvisual Effects of Optical Radiation, for more details). Selfextinguishing lamps usually contain a tungsten filament in place of a portion of nickel wire that will oxidize quickly and separate, extinguishing the electrical arc and turning the lamp off. The lamp is then inoperative and needs to be replaced.

Figure 6-56. Common HID lamp bases (not to scale). ANSI designations are shown. Mercury Lamps.48-52 In mercury lamps, light is produced by the passage of an electric current through mercury vapor. Since mercury has a low vapor pressure at room temperature, and even lower when it is cold, a small amount of more readily ionized argon gas is introduced to facilitate starting. The original arc is struck through the ionization of this argon. Once the arc strikes, its heat begins to vaporize the mercury, and this process continues until all of the mercury is evaporated. The amount of mercury in the lamp essentially determines the final operating pressure, which is 200 to 400 kPa (29 to 58 lb/in.2) in the majority of lamps. The electrodes of mercury lamps usually are made of tungsten, in which the emission material, composed of several metallic oxides, is embedded within the turns of a tungsten coil protected by an outer tungsten coil. The electrodes are heated to the proper electron-emissive temperature by bombardment energy received from the arc. Most mercury lamps are constructed with two envelopes: an inner envelope (arc tube) that contains the arc, and an outer envelope that (1) shields the arc tube from outside drafts and changes in temperature; (2) usually contains a stable, low-pressure gas (generally nitrogen) that prevents oxidation of internal components and also increases the breakdown voltage across the outer bulb parts; (3) provides an inner surface that will accept phosphor coatings; and (4) normally acts as a filter, removing most of the UV radiation produced by the arc. Phosphors placed inside the outer envelope can convert some of this UV energy to light, as in fluorescent lamps. Typically, the mercury lamp's inner envelope (arc tube) is made of fused silica with thin molybdenum ribbons sealed into the ends as current conductors. The outer envelope (bulb) is usually made of hard (borosilicate) glass but also can be of other glasses for special transmission or where pollution and thermal shock are not problems. The essential construction details shown in Figure 6-57 are typical of lamps with fused silica (quartz) inner arc tubes within an outer envelope. Other lamps, such as those for special photochemical application and self-ballasted types, have different constructions.

Figure 6-57. Mercury vapor lamp construction. The pressure at which a mercury lamp operates accounts in large measure for its characteristic spectral power distribution. In general, higher operating pressure tends to shift a larger proportion of emitted radiation into longer wavelengths. At extremely high pressure there is also a tendency to spread the line spectrum into wider bands. Within the visible region the mercury spectrum consists of five principal lines (404.7, 435.8, 546.1, 577, and 579 nm), which result in greenish-blue light at efficacies of 30 to 65 lm/W, excluding ballast losses. While the light source itself appears to be bluish-white, there is a deficiency of longwavelength radiation, especially in low- and medium-pressure lamps, and most objects appear to have distorted colors. Blue, green, and yellow are emphasized; orange and red appear brownish. Clear mercury lamps generally have a CRI value of approximately 15, and are not desirable for use where people will occupy the space. They are, however, quite suited to landscape lighting (see Chapter 21, Exterior Lighting). A significant portion of the energy radiated by the mercury arc is in the UV region. Through the use of phosphor coatings on the inside surface of the outer envelope, some of this UV energy is converted to visible radiation. The most widely used lamps of this type are coated with a vanadate phosphor (4000 K, designation DX) that emits long-wavelength radiation (orange-red); this improves efficacy and color rendering. This phosphor also is blended with others to produce cooler or warmer colors. Figure 6-28 shows the spectral power distributions of a clear lamp and ones using these phosphors. Metal Halide Lamps.53-61 Metal halide lamps are similar in construction to mercury lamps, the major difference being that the metal halide arc tube contains various metal halides in addition to the mercury and argon. When the lamp attains full operating temperature, the metal halides in the arc tube are partially vaporized. When the halide vapors approach the high-temperature central core of the discharge, they are dissociated into the halogen and the metals, with the metals radiating their spectrum. As the halogen and metal atoms move near the cooler arc tube wall by diffusion and convection, they recombine, and the cycle repeats. The use of metal halides inside the arc tube presents two advantages. First, metal halides are more volatile at arc tube operating temperatures than pure metals. This allows the introduction of metals with desirable emission properties into the arc at normal arc tube temperatures. Second, those metals that react chemically with the arc tube can be used in the form of a halide, which does not readily react with fused silica. The efficacy of metal halide lamps is greatly improved over mercury lamps. Commercially available metal halide lamps have efficacies of 75 to 125 lumens/watt (excluding ballast losses). Almost all varieties of white-light metal halide lamps have color rendering properties as good as or superior to phosphorcoated mercury lamps. The radiating metals introduced as halides in these lamps have characteristic emissions that are spectrally selective. Some metals principally produce visible radiation at a single wavelength, while others produce a multitude of discrete wavelengths. Still others provide a continuous spectrum of radiation. In order to obtain a desired spectrum, blends of metal halides are used. Two typical combinations of halides used are scandium and sodium iodides, and dysprosium, holmium, and thulium rare-earth (RE) iodides. Their spectral power distributions are shown in Figure 6-28. Other metals, such as tin, when introduced as halides, radiate as molecules, providing a continuous band spectra across the visible spectrum. The scandium-sodium system, for example, can produce CCTs between 2500 to 5000 K by varying the blend ratio and arc tube operating temperature. The rare-earth system, on the other hand, has a characteristic CCT of approximately 5400 K, which, when augmented by the inclusion of sodium iodide, may be lowered to 4300 K. A rare-earth system augmented with cesium and sodium iodides can achieve a CCT of 3000 K. The rare-earth system provides a somewhat higher general color rendering index than the scandium-sodium system; lithium iodide additions look promising for enhancing the color rendering properties of the scandium-sodium system. Selected colors also can be produced using single elements in the arc tube: sodium for orange, thallium for green, indium for blue, and iron for UV. Luminous efficacy and lamp life tend to be greater for scandium-sodium lamps, but thallium can be used to improve the efficacy of RE lamps. These trade-offs should be considered in selecting a lamp type for each particular application. Metal halide lamps are also available with phosphors applied to the outer envelopes (Figure 1-13). These phosphors lower the CCT of the lamps by approximately 300 K. The main use of the phosphor coating is to create a more diffuse light source. Metal halide lamp construction is similar to that of a mercury lamp (Figure 6-58). One significant design characteristic is that the arc tubes usually are smaller for equivalent wattages. The metal halide arc tube has a white coating applied to the ends to increase vaporization of the metal halides.

Figure 6-58. Metal halide lamp construction. Another design characteristic of metal halide lamps is that the arc tubes often are custom shaped. Most metal halide lamps are life- and lumen-rated in the vertical operating position. For instance, a universal operation lamp has its best performance in the vertical position. When a universal lamp is operated horizontally, the arc bows upward due to convection currents. At the same time, the metal halide pool (which is liquid) moves to the center of the arc tube. The bowed arc moves farther from the metal halides than when the lamp is vertical, causing them to cool. This lowers the vapor pressure of the metal halide chemicals and decreases the concentration of metals in the arc with a resulting loss in light. In addition, the bowed arc moves closer to the top of the arc tube wall, causing its temperature to increase. The higher wall loading on the arc tube material results in a decrease in life rating by approximately 25%. Since many applications require horizontal lamp orientation, a number of arc tube designs have been developed. There are two common configurations for horizontal high output arc tubes as shown in Figure 6-59. The first is a bowed arc tube shaped to follow the natural bowing of the horizontal arc. In this design, the chemicals are confined to the ends of the arc tube as the shape prevents migration. The second design is an asymmetric arc tube with the electrodes lower in the arc tube body such that the arc bows to the center line of the arc tube. Both of these designs provide increased light (approximately 25%) and longer life (approximately 33%) over the universal lamps operated horizontally.

Figure 6-59. Common configurations for horizontal metal halide lamps. Since the horizontal high-output arc tubes are designed to accommodate the upward bow of a horizontally operating arc, the arc tube must be operated horizontally to prevent overheating of the arc tube walls, which dramatically shortens lamp life and increases the probability of violent failures. A special base and socket are always used with horizontal high-output lamps to help ensure proper arc tube orientation (Figure 6-60). Lamp operating position is much less important for mercury and high-pressure sodium lamps.

Figure 6-60. Base and socket configuration for some horizontal high-output metal halide lamps. Some arc tubes have been designed with ovoid shapes (Figure 6-61a). These arc tubes are actually formed in a mold using high-pressure gas. They are commonly referred to as formed body arc tubes. The older style of arc tubes are referred to as pinched body arc tubes (Figure 6-61b). The molding process ensures a highly repeatable and accurate shape for each arc tube. The actual contour and shape of the arc tube gives some excellent benefits in performance. The

walls of the arc tube are contoured to better follow the shape of the arc, thereby allowing for a more uniform thermal profile for the arc tube. This shape also allows the metal halide chemicals to heat up more rapidly than those in the conventional pinched body arc tube. On average, formed body arc tubes warm up three times faster than pinched body arc tubes of the same wattage. Formed body arc tubes have much smaller pinch seal areas. These areas serve to cool the arc tube end chambers and thereby reduce lamp efficacy by lowering the temperature of the metal halide pool. This undesirable cooling is more of a problem in lower-wattage lamps in which the pinch seal area comprises a greater part of the total thermal mass of the arc tube than for the higher wattage lamps.

Figure 6-61. Common arc tubes: (a) ovoid, and (b) pinched body. The smaller pinch seal area has an effect on lamp starting. The older pinch seal designs use a secondary starter electrode that helps to initiate breakdown of the arc tube gases. In formed body arc tubes, there is no room for the secondary electrode. Consequently the arc is initiated by a high-voltage pulse (typically 3,000 V minimum) applied directly across the main electrodes. Devices called ignitors are used to provide these starting pulses. Ignitor starting has been found to have additional performance features. In general, lamps start faster when ignitors are used. They start more reliably, and the fill pressure inside the arc tube can be increased over standard starter electrode systems. This higher fill pressure helps to retard tungsten evaporation from the electrode, which causes lumen depreciation due to arc tube wall darkening. The classic pinched body metal halide lamps that use a starter electrode must also contain a system that provides for either shorting of the starter electrode to the main electrode or opening the starter electrode circuit after the lamps have started. This is required to prevent electrolysis in the fused silica between the starting and operating electrodes, especially when a halide such as sodium iodide is used in the lamp. Failure to short or open the starter electrode circuit will result in very short lamp lives. These starter circuits typically use a bimetal switch. The location and type of switch can restrict the lamp operating position as the bimetal must achieve a certain temperature to function. In some metal halide lamps the electrical connection to the electrode at the dome of the lamps is made by a small nonmagnetic wire remote from the arc tube. This prevents diffusion of sodium through the arc tube by electrolysis caused by a photoelectric effect when the current lead is near the arc tube. Most metal halide lamps above 150 W require a higher open-circuit voltage to start than mercury lamps of corresponding wattage. Therefore, they require specific ballasts. Certain metal halide lamps designs, however, can be operated on some types of mercury ballasts in retrofit situations. Most metal halide lamps use getters to overcome impurities that, if present in the outer jacket of a metal halide lamp in sufficient concentrations, can compromise performance. The predominant problems arise from hydrogen and carbon contamination. Special metal halide lamps are available that automatically extinguish the arc should the outer envelope break or puncture. They can be used in locations where exposure to UV radiation should be avoided.62 Low-wattage metal halide lamps63-65 (below 175 W) come in many varieties for different applications, such as displays, recessed lighting, and track lighting. They produce brilliant white light in a small arc capsule enclosed in a small outer jacket. Such lamps include single-ended lamps with medium or E27 bases (32 to 175 W), single-ended lamps with bipin bases (35 to 150 W), and double-ended lamps with recessed single contact bases (70 to 150 W). Some single-ended lamps use a transparent sleeve surrounding the arc tube called a shroud. A thin-walled shroud is useful as a heat shield because it helps achieve a more uniform arc tube temperature; it also retards sodium loss. A heavy shroud is used in lamps suitable for open luminaires. These shrouds prevent the outer jacket of the lamp from breaking in case of an arc tube violent failure. When relamping an open luminaire it is important to use only open luminaire rated lamps (those with shrouds). To prevent user misapplication the industry has developed unique socket and base combinations for both medium and mogul base lamps. Certain metal halide lamps must be operated in enclosed luminaires designed to contain any hot quartz fragments that might result from an arc tube rupture. Some metal halide lamps do not have a hard glass outer jacket. These lamps can have either no outer jacket or an outer jacket that is made from fused silica that transmits UV energy. In these designs, the luminaire must have a cover glass providing the UV filtration. High-Pressure Sodium Lamps.66 In high-pressure sodium lamps, light is produced by electric current passing through sodium vapor. These lamps are constructed with two envelopes, the inner arc tube being polycrystalline alumina, which is resistant to sodium attack at high temperatures and has a high melting point. Although translucent, this material provides good light transmission (more than 90%). The construction of a typical high-pressure sodium lamp is shown in Figure 6-62. Polycrystalline alumina cannot be fused to metal by melting the alumina without causing the material to crack. Therefore, an intermediate seal is used. Either solder, glass, or metal can be used. Ceramic plugs also can be used to form the intermediate seal. The arc tube contains both xenon as a starting gas and a small quantity of sodium-mercury amalgam, which is partially vaporized when the lamp attains operating temperature. The mercury acts as a buffer gas to raise the gas pressure and operating voltage of the lamp.

Figure 6-62. High-pressure sodium lamp construction. The outer borosilicate glass envelope is evacuated and serves to prevent chemical attack of the arc tube metal parts. It also helps to maintain arc tube temperature by isolating the metal from ambient temperature effects and drafts. Most high-pressure sodium lamps can operate in any position. The operating position has no significant effect on light output. Lamps are also available with diffuse coatings on the inside of the outer bulb to increase source luminous size or reduce source luminance. High-pressure sodium lamps radiate energy across the visible spectrum. This is in contrast to low-pressure sodium lamps, which radiate principally the doublet D lines of sodium at 589 nm. Standard high-pressure sodium lamps, with sodium pressures in the 5 to 10 kPa (40 to 75 Torr) range, typically exhibit color temperatures of 1900 to 2200 K and have a CRI of 22. At higher sodium pressures, above approximately 27 kPa (200 Torr), sodium radiation of the D line is self-absorbed by the gas and is radiated as a continuous spectrum on both sides of the D line. This results in the dark region at 589 nm as shown in the typical spectrum in Figure 6-28. Increasing the sodium pressure increases the CRI to at least 65 at somewhat higher correlated color temperatures; however, life and efficacy are reduced. White high-pressure sodium lamps have been developed with correlated color temperatures of 2700 to 2800 K and a CRI between 70 and 80. Higher-frequency operation is one method of providing white light at reduced sodium pressure. High-pressure sodium lamps have efficacies of 45 to 150 lm/W, depending on the lamp wattage and desired color rendering properties. Because of the small diameter of a high-pressure sodium lamp arc tube, no starting electrode is included as in the mercury lamp. Instead, a high-voltage, highfrequency pulse is provided by an ignitor to start these lamps. Some special high-pressure sodium lamps use a specific starting-gas mixture (a combination of argon and neon that requires a lower starting voltage than either gas alone) and a starting aid inside the outer bulb. These lamps can start and operate on many mercury lamp ballasts. High-pressure sodium lamps are also available with two identical arc tubes contained within the outer bulb. These arc tubes are connected in parallel inside the lamp, but only one arc tube is started with the ignitor pulse. In the event of a momentary power outage, this dual arc tube lamp restrikes immediately when power is restored. Within about one minute, the lamp returns to full light output.

Lamp Designations The current identifying designations of high-intensity discharge lamps generally follow a system that is authorized and administered by ANSI. All designations start with a letter (H for mercury, M for metal halide, S for high-pressure sodium). This is followed by an ANSI-assigned number that identifies the electrical characteristics of the lamp and ballast. After the number there are two letters that identify the size, shape, and finish of the bulb. After this sequence, the manufacturer may add special letters or numbers to indicate information not covered by the standard sequence of the designation, such as lamp wattage or color. An example HID lamp designation is as follows:

Lamp Starting Mercury Lamps. Some special two-electrode mercury lamps, and many photochemical types, require a high open-circuit voltage to ionize the argon gas and permit the arc to strike. In the more common three-electrode lamps an auxiliary starting electrode placed near one of the main electrodes makes it possible to start the lamp at a lower voltage. Here, an electric field is first established between the starting electrode, which is connected to the opposite main electrode through a current limiting resistor, and the adjacent main electrode. This causes an emission of electrons, which develops a local glow discharge and ionizes the starting gas. The arc then starts between the main electrodes. The mercury gradually vaporizes from the heat of the arc and draws current. During this process the arc stream changes from the bluish glow of the argon arc to the blue-green of mercury, increasing greatly in luminance and becoming concentrated along the axis of the tube. At the instant the arc strikes, the lamp voltage is low. Normal operating values are reached after a warmup period of several minutes, during which the voltage rises until the arc attains a stabilization vapor pressure; the mercury is then entirely evaporated. If the arc is extinguished, the lamp will not relight until it is cooled sufficiently to lower the vapor pressure to a point where the arc will restrike with the voltage available. The time from initial starting to full light output at ordinary room temperatures, with no enclosing lighting unit, and also the restriking time (the

cooling time required before the lamp will restart), vary between 3 and 7 min, depending upon the lamp type. Metal Halide Lamps. The method of starting most metal halide lamps above 150 W is through an auxiliary starter electrode similar to that used in mercury lamps. The presence of the metal halides causes the starting voltage to be higher than it needs to be for mercury lamps. Therefore, higher-wattage metal halide lamps generally are not operated on mercury control gear. Sometimes people operate 400-W metal halide lamps on standard mercury ballast. This is not good practice since aged lamps require higher starting voltages, if they start at all. The newer formed body arc tubes do not have room to accommodate the starter electrode in the pinched body arc tube lamps (Figure 6-61b), and therefore a different method of starting is employed. These lamps use an ignitor, a high voltage-low current generating device, as part of the external control circuit. The ignitor provides enough voltage across the main electrodes to initiate an arc. Most lamps below 150 W use an ignitor for starting. A metal halide lamp does not reach full light output immediately but instead must warm up over a period of several minutes (time to reach full light output is longer for higher lamp wattages). During this phase, the color of the discharge changes as the metal halides warm up, evaporate, and incorporate into the arc. Upon full warm up, the lamp color and electrical characteristics stabilize. Since a metal halide arc tube is smaller than that of a mercury lamp of equivalent wattage, it operates at a higher temperature. Hence, the time to cool down is longer and the time to reignite is longer. The hot restrike time in a conventional pinched body arc tube can be 15 min or longer. Lamps that use ignitor starting restrike much faster than the conventional pinched body arc tube designs with starter electrodes due to higher pulse voltages. High-Pressure Sodium Lamps. Since the high-pressure sodium lamp does not contain a starting electrode, a high-voltage, high-frequency pulse is used to ionize the starting gas. Once started, the lamp warms to full light output in approximately 10 min, during which time the color changes. Because the operating pressure of a high-pressure sodium lamp is lower than that of a mercury lamp, the restrike time is shorter. It usually restrikes in less than 1 min and warms up in 3 to 4 min (see also Figure 6-63).

Figure 6-63. Starting and restrike times among different HID lamps.

Lamp Life and Lumen Depreciation Average rated lamp life is defined as that time after which 50% of a large group of lamps are still in operation. The procedure prescribes operating cycles for HID lamps of 11 h on, 1 h off.67 For certain lamp types and applications, criteria other than failure to light may be considered, such as rapid cycling, drastic color change, or significant reduction in lumen output. Lamp life is generally based on the prescribed operating cycle. HID lamp life and lumen maintenance are affected by changes in the operating cycle, however. As a rule of thumb, as the operating period is shortened by 50%, lamp life is reduced by approximately 25%. Contact lamp manufacturers for further information about shorter operating cycles and reduced lamp life. HID lamps usually are rated for initial lumens after 100 h of operation. Figure 6-64 illustrates light losses for three types of 400-W HID lamps over time.

Figure 6-64. Typical lumen maintenance curves for 400-W high-intensity discharge lamps.

Mercury Lamps. General-service mercury lamps have a long average rated life. They usually employ an electrode with a mixture of metal oxides embedded in the turns of tungsten coils from which the electrode is assembled. During the life of the lamp, this emission material is very slowly evaporated, or sputtered, from the electrode and is deposited on the inner surface of the arc tube. This process results first in a white deposit on the inner surface of the arc tube, eventually in a blackening of the arc tube, and ultimately in exhaustion of the emission material in the electrodes and the end of lamp life when the starting voltage exceeds the open-circuit voltage. Metal Halide Lamps. Chemical reaction between the iodine in a metal halide lamp and the emission materials included in mercury lamp electrodes prevents the use of mercury electrodes in a metal halide lamp. Because the electrodes used with metal halide lamps evaporate more rapidly than mercury lamp electrodes, they generally have shorter life ratings. In addition, some metal halide lamps experience strong color changes toward the end of normal life. Where the color appearance of the lamps is critical, the useful life of the lamps ends when the color shift becomes objectionable. High-Pressure Sodium Lamps. High-pressure sodium lamps employ electrodes very similar to those used in mercury lamps. This fact, combined with the smaller diameter of the arc tube, gives high-pressure sodium lamps excellent lumen maintenance.

Figure 6-65. Flicker Index for HID Lamps Operated on Different Ballast Types The life of a high-pressure sodium lamp is limited by a slow rise in operating voltage that occurs over the life of the lamp. This rise is principally caused by arc tube end blackening from electrode sputtering. The blackening absorbs radiation, which heats up the arc tube ends and vaporizes additional sodium amalgam. This increases the arc tube pressure and consequently the arc voltage. Other reasons for arc tube voltage rise are the diffusion of sodium through the arc tube end seals and the removal of sodium from the arc stream by combination with impurities in the arc tube. When the ballast can no longer supply enough voltage to reignite the arc during each electrical half-cycle, the lamp extinguishes. When it cools down, the lamp will again ignite and warm up until the arc voltage rises so that the ballast cannot support the arc. This cycling process occurs until the lamp is replaced.

Effect of Ambient Temperature The light output of a typical double-envelope HID arc tube is little affected by the ambient temperature. These lamps are generally satisfactory for temperatures down to − 29°C (− 20°F) or lower. On the other hand, single-envelope lamps, intended primarily for use as UV sources, are critically affected by low temperatures, particularly if the surrounding air is moving. They are not considered suitable for use below 0°C (32°F) without special protection, since they do not give full output. Ambient temperature affects the striking voltage of all discharge lamps, and in some cases higher starting voltages for indoor use are recommended for roadway and floodlighting installations in cold climates. Ballasts for roadway lighting service and other low-temperature applications are designed to provide the necessary voltage to start and operate each particular lamp at temperatures as low as − 29°C (− 20°F). Recommendations for starting voltages have been developed by ANSI.68

Lamp Operating Temperature Excessive envelope and base temperatures may cause failures or unsatisfactory performance due to softening of the glass, damage to the arc tube by moisture driven out of the outer envelope, softening of the basing cement or solder, or corrosion of the base, socket, or lead-in wires. Maximum bulb and base temperatures are prescribed by various standards associations. The use of reflecting equipment that concentrates heat and energy on either the inner arc tube or the outer envelope should be avoided. In the case of metal halide and high-pressure sodium lamps in which all the material is not vaporized, concentrated heat on the arc tube can affect the color of illumination as well as electrical characteristics and lamp life.

Flicker and Stroboscopic Effect The light output of all HID lamps varies to some degree with cyclic changes of the line voltage. This flicker depends on the lamp type and the ballast circuit. Flicker can be an important consideration for HID lamps. In many lighting applications the stroboscopic effect from HID sources is not a problem. It can, however, be annoying to spectators in games such as tennis or Ping-Pong. Operators of rotating machinery can find it distracting. To minimize the stroboscopic effect, systems with a flicker index of 0.1 or less are suggested, or luminaires can be wired alternately on different phases of a three-phase system. Figure 6-65 illustrates the variation in flicker index for mercury, metal halide, and high-pressure sodium lamps for several ballast types operated at 60 Hz. The flicker index is considerably higher in 50-Hz power systems. The flicker effect can be effectively eliminated by using electronic ballasts having high-frequency or rectangular wave characteristics. In addition to the above, some flicker sometimes can be seen from the end of the lamp when viewed peripherally by the retina (see Chapter 3, Vision and

Perception). This flicker is a result of the arc being initiated at alternate electrodes during positive and negative half cycles, and has a frequency that is equal to line frequency. It is also eliminated in high-frequency operation.

Auxiliary Equipment HID lamps have negative volt-ampere characteristics, and therefore a current-limiting device, usually in the form of a transformer and reactor ballast, must be provided to prevent excessive lamp and line currents. The lamps are operated on either multiple or series circuits. Figure 6-66 gives schematic diagrams of several typical ballast types. Figure 6-67 summarizes the characteristics of the most common combinations of HID lamps and ballasts. A distinction must be made between lag circuit and lead circuit ballasts. The lamp current control element of a lag circuit ballast consists of an inductive reactance in series with the lamp. The current control element in lead circuit ballasts consists of both inductive and capacitive reactances in series with the lamp; however, the net reactance of such a circuit is capacitive in mercury and metal halide ballasts, and inductive in high-pressure sodium ballasts.

Figure 6-66. Typical circuits for operating high-intensity discharge lamps.

Figure 6-67. Characteristics of Common Combinations of HID Lamps and Ballasts There are a number of ballasts in use for operating mercury lamps. Wattage losses in ballasts are usually in the order of 5 to 15% of lamp wattage. Lag Reactor. The simplest lag circuit ballast is a reactor consisting of a single coil wound on an iron core placed in series with the lamp. The only function of the reactor is to limit the current delivered to the lamp. Such a reactor can be used only when the line voltage is within the specified lamp starting voltage range. The power factor of this circuit is approximately 50% lagging; this is commonly referred to as normal or low power factor. The line current under starting conditions is approximately 50% higher than normal operating current; therefore, it is recommended that supply wiring be sized for approximately twice the normal operating current. High-power-factor versions are available where a capacitor is installed in the circuit to increase the power factor of the system to better than 90%. This is generally the preferred system, since it also reduces the input current under starting and operating conditions almost 50% below that of the low-power-factor system, allowing full utilization of the circuit. Since a lag reactor performs only the function of current control, it is the smallest, most economical, and most efficient ballast. However, it has shortcomings that should be considered in application. The reactor provides little regulation for fluctuations in line voltage; for example, a 3% change in line voltage can cause a 6% change in lamp wattage. Therefore, the reactor is not recommended where line fluctuations exceed 5%. High-Reactance Autotransformer. Where the line voltage is below or above the specified lamp starting voltage range, a transformer is used in conjunction with the reactor to provide proper starting voltage. This normally is accomplished with the combination of primary and secondary coils forming a one-piece

single high-reactance autotransformer. The power factor of this circuit is approximately 50% lagging and has the same advantages and shortcomings as the normal power factor of a reactor lag circuit. High-power-factor versions are available in which a capacitor is installed in the circuit to increase the power factor of the system to better than 90%. The effect on input current is the same as in the high-power-factor reactor. Regulation and lamp performance are unchanged. Constant-Wattage Autotransformer (CWA). This type of lead circuit ballast is the most widely used in mercury lighting systems. It consists of a highreactance autotransformer with a capacitor in series with the lamp. The capacitor allows the lamp to operate with better wattage stability when the voltage on the branch circuit fluctuates. This ballast is used when line voltage is expected to vary by more than 5%. For example, a 10% change in line voltage would result in only a 5% change in lamp wattage. Other advantages with the CWA ballast are high power factor, low line extinguishing voltage, and line starting currents that are lower than normal line currents. The CWA ballasts allow for maximum loading on branch circuits and provide an economical and efficient mercury lighting system. The capacitor used with the CWA ballast performs an important ballasting function, as in all lead-type circuits. The capacitor used in lag-type high-power-factor reactor and high-power-factor autotransformer ballasts is purely a power factor correction component and has no ballasting function. Constant Wattage (CW). This type of ballast, also referred to as regulated or stabilized, has operating characteristics similar to the CWA. The light output and wattage vary less than 2% with up to a 13% change in line voltage. The CW ballast, like the CWA ballast, uses a lead circuit; it differs in that the lamp circuit is completely isolated from the primary winding. It also has the same advantages as the CWA ballast, such as high power factor, low line extinguishing voltage, and low line starting currents. Two-Lamp Lead-Lag Circuit. The lead-lag ballast design approach is commonly used to operate two 250-, 400-, or 1000-W mercury or metal halide lamps in two independent circuits. A current-limiting reactor operates one lamp, and a combination reactor and capacitor connected in series operates the second. The lamps operate independently, so that the failure of one has no effect on the other. The input current of the combination of capacitors and reactors is lower than the sum of the two individual operating currents. These elements provide a high power factor and reduce flicker. This circuit can be used only when the line voltage is within the specified lamp starting voltage range. It is the most economical two-lamp system with regulation similar to the normal-power-factor reactor and autotransformer ballasts. A lag reactor may be used in one luminaire and a lead reactor in the next luminaire. An equal number of each in a branch circuit will result in a high branch circuit power factor. Two-Lamp Series (Isolated) Constant Wattage. This circuit is essentially the same as single-lamp constant wattage, except that it operates two lamps in series. The most effective use of this circuit is in applications where the ambient temperature is − 18°C (0°F) or above. It is most popular for indoor 400-W applications. Constant-Current Series Regulators. Mercury lamps also are operated in series on constant-current series regulators. The most commonly used method employs a current transformer for each lamp. It differs in design from the more common multiple type of ballast. The usual design is a two-winding transformer as illustrated in Figure 6-66e. Since the series circuit regulator reactance limits the current in the circuit, the individual lamp current transformer is not designed to limit current, but rather to transform it from the regulator secondary current to the rated lamp current. In addition, the transformer is made to limit the secondary open-circuit voltage so that no cutout is necessary when a lamp fails. Series transformers are available for the more popular lamps to operate from either 6.6-, 7.5- or 20-A series circuits and can be operated on all types of constant-current transformers. These circuits usually are satisfactory for metal halide lamps and high-pressure sodium lamps designed for operation on reactor-type mercury ballasts. Two-Level Mercury Ballasts. Two-level operation of mercury lamps can be accomplished by switching capacitors on lead circuit ballasts. Such ballasts that operate 125-, 250- and 400-W mercury lamps at two levels are available. For example, a 400-W mercury lamp may be operated at either 400 or 300 W by switching leads at the ballast. These two-level mercury ballasts are used for energy saving. Similar designs are available for high-pressure sodium and metal halide lamps. Both lamp manufacturer and ballast manufacturer should be contacted for specific information. This control technique is presently limited to horizontal operation above 10°C (50°F). The warmup time is 50% longer on low level. Metal Halide Lamps. Metal halide lamps operate well on many different types of control gear. The control gear selection depends on the operating environment of the lamp and the degree of voltage regulation provided by the local utility. In the United States, the common type of control gear for lamps over 150 W is the lead peaked autotransformer. This control gear provides moderately good voltage regulation, yielding a change in lamp wattage of 7 to 10% for a line voltage change of 10%. The lead peaked autotransformer performs much like the CWA ballast and has similar operating features. In Europe and many other countries where high open-circuit voltages are available from the mains, lag-reactor ballasts plus an ignitor are more commonly used. Where voltage regulation is good, the use of lag-reactor ballasts can save significant energy over the more common multi-tap lead peaked autotransformer. For metal halide lamps below 175 W, the most common control gear is the lag reactor or high-reactance autotransformer. Power correcting capacitors typically are employed. Using the lag reactor ballast to regulate lamp wattage causes the least line voltage variations. A 5% change in line voltage can result in a 12% change in lamp wattage. Long-term operation of lamps under high line conditions shortens lamp life. For starting, an ignitor is used that provides a high voltagelow current pulse of between of between 3 and 5 kV. There are three types of ignitors in use today. The most common is the impulser or parallel ignitor, which uses a ballast winding as the ignitor's pulse transformer. Another type is the superimposed or series ignitor, which contains a pulse transformer that is independent of the ballast windings. Finally, there is the two-wire ignitor, which provides a lower pulse voltage directly across the lamp leads. High-Pressure Sodium Lamps. Unlike mercury and metal halide lamps, which exhibit relatively constant lamp voltage with changes in lamp wattage, the high-pressure sodium lamp voltage varies with lamp wattage. Therefore, operating parameters for maximum and minimum permissible lamp wattage and voltage have been established.66 The latest ANSI recommendations for high-pressure sodium lamps should be followed. Figure 6-68 shows the lamp voltage and wattage limits for 400-W high-pressure sodium lamps. Lag or Reactor Ballast. This ballast type is similar to the mercury lamp reactor ballast. It is a simple reactor in series with the lamp, designed to keep the operating characteristics within the trapezoid. A starting circuit is incorporated to provide the starting pulse. Step-up or step-down transformers are provided where necessary to match the line voltage. In most cases, a power-factor-correcting capacitor is placed across the line or across a capacitor winding on the ballast primary. This type of ballast usually provides good wattage regulation for variations in lamp voltage, but rather poor regulation for variations in line voltage. It is the least costly ballast with the lowest power loss among ballasts for high-pressure sodium lamps.

Figure 6-68. Wattage and voltage limits for 400-W high pressure sodium lamps. Magnetic Regulator or Constant-Wattage Ballast. This ballast consists essentially of a voltage-regulating section that feeds a current-limiting reactor and the pulse starting circuit. It provides good wattage regulation for changes in line voltage, as a result of the voltage-regulating section, and good regulation for changes in lamp voltage, which is the main characteristic of the reactor ballast. The magnetic regulator is a high-cost ballast, having the greatest power losses, but it generally provides good wattage regulation under all conditions of line and lamp voltage. A power-factor-correcting capacitor usually is included. It should be noted that this circuit differs from constant-wattage mercury ballast circuits. Lead Circuit Ballast. This circuit is similar to that of the CWA mercury-lamp ballast. It operates with a combination of inductance and capacitance in series with the lamp. It differs in design from the CWA mercury-lamp ballast in that it does not maintain a constant lamp current but rather decreases the current as the lamp voltage increases, so as to keep the lamp operating wattage within the trapezoid limits. This ballast type provides wattage regulation for changes in both line voltage and lamp wattage. For a change of no greater than 10% in line voltage it maintains the lamp wattage within the trapezoid. It is intermediate in cost and power losses. Ignitors. Ignitors are used in the ballast circuit for most high-pressure sodium lamps, some metal halide lamps and some specialty arc lamps. The ignitor starts cold lamps by first providing a high enough voltage for ionization of the gas to produce a glow discharge. To complete the starting process, enough power must then be provided by the starting pulses to sustain an arc through a glow-to-arc transition. The range of pulse voltages to start cold lamps is 1 to 5 kV, usually provided by an electronic resonant circuit which applies multiple pulses to the lamp when the circuit is energized. The circuit turns itself off after the lamp starts by sensing the reduction in open-circuit voltage or, with some ignitors, after a fixed period of time. Instant restarting of hot lamps is accomplished by increased ignition voltage. Voltage pulses of 10 to 70 kV are required by most HID lamps, and these are provided by resonant circuits. To halve the voltage to ground values, ignitor circuits are available to apply opposing pulses simultaneously to the ends of the lamp. Most instant-restart lamps are of double-ended construction to minimize arc-over between lead wires, internal supports, or base contacts. These highvoltage starting pulses normally are applied in one or several short bursts, using the open-circuit voltage reduction on restart to turn off the ignitor.

Dimming of High-Intensity Discharge Lamps Although HID lamps are optimized to operate at full power, some energy savings can be obtained through dimming. In energy management applications, savings of 50% or more might be obtained where available daylight is used with a photosensor and dimming control system. Daylight tends to compensate for the color changes of dimmed HID lamps. Additional controls for HID lighting can be employed for lumen maintenance and with time-of-day and demand reduction programs as administered by a computer or a simple time clock. Simple manual control for an HID dimming system can provide flexibility in multipurpose room applications and improved efficiency by tuning the light output for a specific task (see Chapter 27, Lighting Controls). Some light sources are more suitable for dimming than others. Therefore, when considering a particular HID light source for dimming, it is suggested that the manufacturers of the lamp and the dimming system be contacted for information or performance characteristics in the dimmed state. In some cases, the lamp manufacturer's warranty is limited when dimming. In a properly designed dimming system, however, lamp life is unaffected by dimming. The slow warmup and hot restrike delay, which are characteristic of HID sources, also apply to dimming. HID lamps respond to changes in dimmer settings much more slowly than incandescent or fluorescent sources; delays between minimum and maximum light output vary 3 to 10 min. Instantaneous dimming is available, however, over a limited range for some lamps. HID lamps should be started at full power and the dimming delayed until the lamp is fully warmed up. Properly designed dimming systems ensure this occurs. In addition to speed, the range of response is not comparable to that of incandescent or fluorescent dimming; however, in most cases the lamp efficacy and color are reasonably good down to 50% dimming or less. While not well suited to dramatic lighting or theatrical effects, they are quite satisfactory for many energy management applications. In these applications, the slow response of HID lamps provides additional system stability and minimal occupant distraction. Typical curves of lumen output versus input power are shown in Figure 6-69. These curves describe general trends; the dimming system manufacturer should be consulted for more details. Clear mercury lamps change very little in color from 100 to 25% light output; the blue-green color that is characteristic of clear mercury sources is present at all dimmer settings (Figure 6-70). Color-improved mercury lamps generally perform well down to approximately 30% light output. The color appearance, color consistency from lamp to lamp, and color rendering of some clear, low-wattage metal halide lamps can begin to change at 80% lumen output. For higher wattages, the color of the clear lamps begins to change at approximately 60% light output, where a blue-green color (characteristic of mercury vapor) starts to appear. The effect is somewhat less with phosphor-coated lamps. The color appearance of typical high-pressure sodium lamps does not appreciably change until approximately 50% light output. Below 50%, a strong yellow color, characteristic of low-pressure sodium, begins to prevail. In the case of the higher-color-rendering high-pressure lamps, the lamp manufacturer should be consulted on dimming performance.

Figure 6-69. Lumen output vs. power input for high-intensity discharge lamps: (a) mercury vapor, (b) metal halide, and (c) high-pressure sodium. The dotted lines represent operating areas where significant color changes in the lamp occur.

Figure 6-70. Correlated color temperature and CRI shift with dimming. Clear metal halide lamps can experience a shift in color temperature of over 1000 K and a drop of 35% in CRI when dimmed to 50% of rated output. Phosphor-coated lamps are much less vulnerable to this color shift.

Self-Ballasted Lamps Self-ballasted mercury lamps are available in various wattages. These lamps have a mercury vapor arc tube in series with a current-limiting tungsten filament. In some types, phosphors coated on the outer envelope provide additional color improvement. The overall efficacy is lower than that of other mercury lamps because of the resistive losses of the tungsten filament. As the name denotes, these lamps do not require an auxiliary ballast.

SHORT-ARC LAMPS Short-arc or compact-arc lamps characteristically provide a source of very high luminance. They are primarily used in searchlights, projectors, display systems and optical instruments (e.g., spectrophotometers and recording instruments) and for simulation of solar radiation. They also can be used as sources of modulated light, generated through current modulation. Short-arc lamps are high-pressure gas discharge lamps characterized by an electrode-stabilized arc that is short compared with the size of the envelope. Depending on rated wattage and intended application, their arc length may be from about 0.01 to 0.47 in. (0.3 to 12 mm). These arcs have the highest luminance and radiance of any continuously operating light source and are the closest to being a true point source.69-82 Some typical short-arc lamps are shown in Figures 6-71 and 6-72. These lamps have optically clear fused silica (quartz) bulbs of spherical or ellipsoidal shape with two diametrically opposite seals. Four types of seal are used in short-arc lamps. The graded seal and the molybdenum foil seal are current-carrying seals, while the molybdenum-and-Kovar cup seal and the elastomer (mechanical) seal are separated from the current conductor by a cup or a flange. For applications requiring ozone-free operation, the lamp envelopes are fabricated from quartz, which does not transmit wavelengths below 210 nm. Most short-arc lamps are designed for dc operation. The better arc stability and substantially longer life of the dc lamps have limited the use of ac short-arc lamps to special applications.

Figure 6-71. Typical short-arc lamps: (a) low-wattage mercury-argon lamps (100-W at left, 200-W at right), and (b) medium-wattage xenon lamps (from left: 1.6, 2.2, 4.2, 3.0 kW).

Figure 6-72. Typical high-power xenon compact arc lamps with liquid-cooled electrodes: (a) 30-kW lamp for solar simulators (principal operating position is vertical), and (b) 20-kW lamp for military searchlights (principal operating position is horizontal).

Starting of Short-Arc Lamps Like most vapor discharge lamps, short-arc lamps require auxiliary devices to start the arc and limit the current. For ac lamps, either resistive or inductive ballasts are used. Direct-current lamps are best operated from specifically designed power systems that provide, with good efficiency, the high-voltage pulses (up to 50 kV) required to break down the gap between the electrodes, ionize the gas, and heat the cathode tip to thermionic emitting temperatures. Further, they provide enough open-circuit dc voltage to assure the transition from the low-current, high-voltage spark discharge initiated by the starter to the high-current, low-voltage arc. With a properly designed system, a short-arc lamp will start within a fraction of a second. Many power supplies are regulated so that lamp operation is independent of line voltage fluctuation. Four basic types of sensors for power regulation are presently in use: current, voltage, power, and optical regulators. The type of power system used depends on the specifics of the application.

Mercury and Mercury-Xenon Lamps69,70,72,76 To facilitate lamp starting, short-arc mercury lamps contain argon or another rare gas at pressures of 1 to 4 kPa (0.15 to 0.58 lb/in.2), the same as standard mercury lamps. After the initial arc is struck, the lamp gradually warms up, and the voltage increases and stabilizes as the mercury is completely vaporized. Mercury lamps require several minutes to achieve full operating pressure and light output. This warmup time is reduced by approximately 50% if xenon at greater than atmospheric pressure is added to the mercury. Lamps with this type of fill are known as mercury-xenon short-arc or compact-arc lamps. The spectral power distribution in the visible region is essentially the same for both types, consisting mainly of the five mercury lines and some continuum due to the high operating pressure (Figure 6-73). The luminous efficacy of these lamps is approximately 50 lm/W at 1000 W and approximately 55 lm/W at 5000 W. Mercury and mercury-xenon lamps are available from 30 to 7000 W and are usually operated in the vertical position.

Xenon Lamps71,74-82 Xenon short-arc lamp are filled with several atmospheres of xenon gas. They reach 80% of the final output immediately. The CCT of the arc is approximately 5000 K. The spectrum is continuous from the UV through the visible into the infrared (Figure 6-74). Xenon lamps exhibit strong emission lines in the near infrared between 800 and 1000 nm and some weak lines in the short-wavelength visible region. Xenon short-arc lamps range from 5 to 32,000 W, and they are available for operation in vertical or horizontal positions. Lamps designed for operation above 10 kW typically require liquid cooling of the electrodes. The luminous efficacy of the xenon short-arc lamp is approximately 30 lm/W at 1000 W, 45 lm/W at 5000 W, and over 50 lm/W at 20 kW and above.

Figure 6-73. Spectral power distribution of a 2.5-kW mercury-xenon lamp from 240 to 2500 nm.

Figure 6-74. Spectral power distribution of a 2.2-kW xenon lamp from 200 to 2500 nm.

Ceramic Reflector Short-Arc Xenon Lamps Ceramic reflector xenon (CRX) lamps combine the basic technology of short-arc lamps and an internal reflector that focuses the arc's energy through a sapphire window in the ceramic-to-metal housing. The window transmits energy in both the IR and UV regions of the spectrum. These devices have the advantages of increased output, safety, and ease of handling and installation. They also obviate some peripheral equipment such as optical focusing components needed in many applications where short-arc lamps are used. CRX lamps are available with inputs from 175 to 1000 W. Their spectral power distributions can be varied for special applications by altering the reflector material or its coating.

Lamp Operating Enclosure Short-arc mercury, mercury-xenon, and xenon lamps are under considerable pressure during operation (up to 5000 kPa [726 lb/in.2] for small lamps and approximately 1000 kPa [145 lb/in.2] for large ones) and therefore must be enclosed during operation. In addition, precaution must be taken to ensure protection from the powerful UV radiation emitted from these lamps. See Chapter 5, Nonvisual Effects of Optical Radiation. In general, short-arc lamps up to approximately 1 kW are designed to operate with convection cooling. Special ventilation is not required unless critical components of the lamps are subjected to excessive temperatures, caused by confined enclosures, excessive ambient temperatures, or infrared radiation. For safety during shipment, storage, or handling of xenon or mercury-xenon lamps, special protection cases are provided. These cases are made of metal or plastic and are so arranged around the bulb that the lamp can be electrically connected without removing the case. The case should not be removed until immediately before the lamp is energized.

Compact-Source Metal Halide Lamps Compact-source, or medium-arc, metal halide lamps83,84 are based on a combination of the short-arc lamp and the metal halide lamp technology. Their arc discharge is predominantly electrode stabilized and operates between tungsten electrodes spaced 2.5 to 35 mm (0.1 to 1.4 in.) apart in ellipsoidal or almost spherical quartz bulbs. They are filled with mercury and argon as basic elements for starting the arc, and, as in some standard metal halide lamps, rare-earth metal iodides and bromides are added in order to obtain a broad spectral power distribution. Both a high luminous efficacy and excellent color rendering with a

correlated color temperature close to that of daylight are achieved, together with a small source size. These lamps are available in various single-ended and double-ended constructions typically ranging from 70 to 1800 W. In special cases, higher wattages (up to 18,000 W) may be achieved. They typically are designed to operate on alternating current and require unique power supplies or ballasting equipment with highvoltage starting devices. Most lamps can be instantly restarted when hot. Some lamps are available with the arc tubes mounted in integral ellipsoidal reflectors that can focus the light through small apertures or fiber-optic bundles. Other lamps are available in PAR configurations for applications that require concentrated beams. Compact-source lamps are used for motion picture and television lighting, outdoor location lighting, theatrical lighting, sports lighting, fiber-optic illuminators, liquid crystal displays (LCD), and video projectors. If used in projectors, careful attention must be given to the modulation if the lamp output and the projector's shutter speed (frame rate) to avoid unintentional stroboscopic effects.

MISCELLANEOUS DISCHARGE LAMPS Low-Pressure Sodium Lamps In low-pressure sodium discharge lamps, the arc is carried through vaporized sodium. The light produced by the low-pressure sodium arc is almost monochromatic, consisting of a double line at 589.0 and 589.6 nm. The starting gas is neon with small additions of argon, xenon, or helium. In order to obtain the maximum efficacy of the conversion of the electrical input to the arc discharge into light, the vapor pressure of the sodium must be approximately of 0.7 Pa (1/10−4 lb/in.2), which corresponds to an arc tube bulb wall temperature of approximately 260°C (500°F). Any appreciable deviation from this pressure degrades the lamp efficacy. To maintain the operating temperature for this pressure, the arc tube is normally enclosed in a vacuum flask or in an outer bulb at high vacuum. The run-up time to full light output is 7 to 15 min (Figure 6-75). When first started, the light output is the characteristic red of the neon discharge, and this gradually gives way to the characteristic yellow as the sodium is vaporized. The hot re-ignition is good, and most low-pressure sodium lamps restart immediately after interruption of the power supply.

Figure 6-75. Low pressure sodium lamp performance during starting. Efficacy. Low current density is vital to efficient generation of light. High densities result in excitation of atoms to higher energy levels and thus a loss of efficacy in converting electricity to light. Thermal insulation increases efficacies to above 180 lm/W for the 180-W U-type low-pressure sodium lamp, or approximately 150 lm/W including ballast losses. The thermal insulation consists of a light-transparent, IR-reflecting layer on the inside of the outer envelope. It is generally made of indium oxide; previously tin oxide and internal glass sleeves were used. Construction. There are two types of low-pressure sodium lamps: the linear and the hairpin, or U tube. The linear lamp has a double-ended arc tube, similar to a fluorescent lamp, with preheat electrodes sealed into each end. Its arc tube, made of a special sodium-resistant glass, is sealed into an outer vacuum jacket with a medium bipin base at each end. The hairpin type has the arc tube double back on itself, with its limbs very close together (Figure 6-76). Two versions are available, based on different approaches to maintaining even distribution of sodium in the arc tube throughout life. Excess metallic sodium condenses at the coolest part of the lamp, generally at the bend of the arc tube. If not controlled, very low sodium concentrations will produce a neon-argon arc in the tube. This is known as "operating bare."

Figure 6-76. Construction of low pressure sodium lamps (U-tube or hairpin type). One hairpin design provides dimples in the outer surface of the arc tube; these serve as alternative cool points for the metallic sodium condensation. The dimples also inhibit migration of sodium due to vibration or gravitational effects. The alternative design uses a graded heat-reflecting film along the inside of the outer envelope, with the greatest amount of reflected heat at the bend. Auxiliary Equipment. The low-pressure sodium arc, in common with all discharge lamps, has a negative volt-ampere characteristic. A current-limiting device, usually a transformer and reactor ballast, must be provided to prevent excessive lamp and line currents. High-power-factor autotransformer ballasts are most commonly used. The required lamp starting voltages ranging between 400 and 550 V. A capacitor wired in parallel on the primary side increases the power factor to 90% or better. On this type of ballast, lamp regulation is excellent: lamp wattage and lumen output remain within 5% for a varying line voltage range of 20%. Constant-wattage ballast designs are also available.

Glow Lamps These are low-wattage, long-life lamps designed primarily for use as indicator or pilot lamps, night lights, location markers, and circuit elements. They range from 0.06 to 3 W and have an efficacy of approximately 0.3 lm/W. A group of typical glow lamps is shown in Figure 6-77. These emit light having the spectral character of the gas with which they are filled. The most commonly used gas is neon, having a characteristic orange color. The glow is confined to the negative electrode. Glow lamps have a critical starting voltage, below which they are, in effect, an open circuit.

Figure 6-77. Typical glow lamps with ANSI numbers (old trade numbers). Like other discharge lamps, glow lamps require a current-limiting resistance in series. Glow lamps with screw bases have this resistor built into the base, while for unbased lamps or lamps with bayonet bases a resistor of the proper value must be employed external to the lamps. Glow lamps filled with an argon mixture rather than neon radiate chiefly in the near UV region around 360 nm and are therefore used mainly to excite fluorescence in minerals and other materials as well as for some photographic applications.

Zirconium Concentrated Arc Lamps, Enclosed Type These lamps use a direct-current arc constituting a concentrated point source of light of high luminance, up to 45 million cd/m2. They are made with permanently fixed electrodes sealed into an argon-filled glass bulb. The light source is a small spot, 0.13 to 2.8 mm (0.005 to 0.11 in.) in diameter (depending on the lamp wattage), which forms on the end of a zirconium oxide-filled tantalum tube that serves as the cathode. The spectral power distribution is similar to that of a blackbody with a correlated color temperature of 3200 K. These lamps produce a candela distribution characterized by the cosine law. They require

special circuits that generate a high-voltage pulse for starting and a well-filtered and ballasted operating current. Suitable power supplies are recommended by the manufacturer. Figure 6-78 illustrates various examples of side- and end-emission lamps.

Figure 6-78. Side and end-emission concentrated arc lamps.

Pulsed Xenon Arc (PXA) Lamps These are ac xenon lamps with two active electrodes (a polarized xenon lamp has current flowing in only one direction and one active electrode). A switching reactor in series with the low-pressure lamp forces 50 to 100 peak amperes (120 pulses per second) through the lamp. The reactor also supplies a continuous current of 2 to 3 A to keep the lamp operating between pulses. The spectrum produced is characteristic of xenon, typically 6000 K. PXA lamps are available in linear and helical types. The efficacy of these sources is approximately 35 to 40 lm/W. Available lamp wattages range from 300 to 8000 W, and forced-air cooling is required during operation. PXA lamps are used in the graphic arts industry for applications requiring instant start; high-intensity, stable light output; and daylight-quality color temperature.

Flashtubes These light sources are designed to produce high-intensity flashes of extremely short duration. They are primarily used for photography; viewing and timing of reciprocating and rotating machinery; airport approach lighting systems, including navigation aids, obstruction marking, and warning and emergency lights; laser pumping; and entertainment applications. A conventional flashtube consists of a transparent tubular envelope of glass or fused silica (quartz) that has its main discharge electrodes internally located near the extremities and usually has an external electrode of wrapped wire for triggering. It generally contains very pure xenon gas at a pressure below atmospheric, usually in the range of 25 to 80 kPa (3.63 to 11.6 lb/in.2). Sometimes other gases such as argon, hydrogen, and krypton are added to the xenon to obtain different spectral power distributions or different electrical, thermal, and deionization characteristics. With a voltage applied across its main electrodes, the tube acts as a high impedance or open circuit until a trigger pulse ionizes the gas within the tube. A trigger pulse induces ionization and thereby causes the xenon gas to become conductive. A discharge then occurs between the main electrodes, whereupon the gas becomes highly luminescent. A xenon flashtube converts up to 35% of the input energy to light. The luminous efficacy ranges from 30 to 60 lm/W. The spectral quality is close to that of daylight, having a correlated color temperature of approximately 6000 K, so that the radiation encompasses the entire visible spectrum and extends into the UV and near IR (Figure 6-79). Flashtubes are available in many sizes and shapes to suit the user and the type of optical system employed. The most common types are straight (linear), wound (helix), and U shape. Other configurations are available for special applications. Figure 6-80 shows some typical commercially available flashtubes.

Figure 6-79. Spectral power distribution of a typical xenon-filled flashtube for two different discharge conditions: (a) high voltage, low capacitance (solid line), and (b) low voltage, high capacitance (dashed line).

Figure 6-80. Typical flashtubes. Energy and Life. For single-flash operation the limit to the amount of energy that can be consumed depends on the desired tube life measured in useful flashes. This life is affected by the rate of envelope wall blackening and destruction of the tube or its parts. Flashtubes designed for very high loading have envelopes made of fused silica, which can withstand high thermal shock. The peak power encountered during a discharge produces a thermal shock that could shatter the envelope; hence, to maximize the energy per flash the thermal shock must be limited. This can be done by reducing the peak current, which also lengthens the flash duration. To limit peak current and thermal shock as well as control the pulse duration, inductance is added in series within the discharge loop. Normally, the life expectancy of a flashtube can be related to the percentage of explosion energy expended by a flash in a particular application. The explosion energy is defined by manufacturers as the energy level at a given flash duration that will cause the tube to fail within ten flashes, usually by disintegration of the envelope. The life can be approximated as follows:

Limits of Power Input. The average power input is a product of the energy per flash and the flash rate. The maximum power that any flashtube can dissipate is determined by the envelope area, the type of envelope material, and the method of cooling, such as free air convection, forced air convection, or use of a liquid coolant. For fused silica envelopes the maximum input power can be approximated as follows: free air convection, 5 W/cm2 (32.3 W/in.2); forced air convection, 40 W/cm2 (258 W/in.2); liquid cooling, 200 W/cm2 (1290 W/in.2). Energy Storage Banks. The electrical energy that subsequently is discharged through the flashtube to produce light is stored in a capacitor bank. This bank must be capable of rapid discharge into a very low impedance load. Therefore, it must have a rather low inductance as well as a very low equivalent series resistance. It must also be capable of storing energy at a high voltage without significant leakage. Typical voltages vary from about 300 to 4000 V. Current banks use aluminum electrolytic, paper-oil or metalized paper capacitors designed specifically for energy storage applications. All are highly efficient in delivering energy to the flashtube. The type selected depends upon the voltage, temperature, and life, as well as size and weight limitations. Electronic Circuitry. In addition to discharge circuitry, a conventional xenon flash system has a charging circuit and a trigger circuit (Figure 6-81). The charging circuit accepts primary electrical power at low voltage, transforms and rectifies it to higher voltage, and applies it to the capacitor bank, where it is stored as potential energy. The luminance of the flashtube depends upon its loading, which in turn depends on the capacitance of the energy storage capacitor and the voltage across it, in accordance with the formula:

where C = capacitance in µF V = voltage across the tube (and capacitor) in kV. The trigger circuit used for producing the high-voltage ionizing pulse consists of a low-energy capacitor discharge system driving a pulse transformer. The pulse transformer establishes an electric field that starts the ionization process and causes the gas to conduct. This pulse usually is applied to the external trigger wire (external electrode), but in some applications it is applied across the main discharge electrodes by a pulse transformer with a very low secondary impedance in series with the flashtube discharge circuit.

Figure 6-81. Basic elements of a typical flashtube power supply.

By varying the voltage to the capacitor bank and hence its capacitance, and by the insertion of inductance in the discharge circuit, it is possible to vary both the light output and the flash duration of the system. The flash duration is dependent on the value of the capacitor, the inductance of the discharge circuit, and the effective impedance of the flashtube. Although the flashtube is a nonlinear circuit element, its effective impedance can be approximated according to the formula:

where ρ = plasma impedance in Ω × cm, L = arc length in cm, A = cross-sectional area of the arc in cm2. At current densities encountered in usual practice, ρ has a value of approximately 0.02 Ω-cm. Flashtubes with their associated circuitry can be designed to operate with flash energies from fractional watt-seconds to 20,000 watt-seconds, with durations from approximately one microsecond to many milliseconds, and with repetition rates from a single flash up to 1000 flashes per second. Even higher repetition rates can be attained with special circuitry and flashtube design.

Linear-Arc Lamps Linear-arc quartz envelope lamps are available for both continuous wave and pulsed operation. Lamps operated in the pulsed mode are discussed above under flashtubes. Forced-air-cooled long-arc xenon lamps are made with arc lengths up to 1.2 m (4 ft), bore diameters up to 12 mm (0.47 in.), and wattages up to 6 kW. These lamps are used for special illumination requirements and solar simulation and have an efficacy of approximately 30 lm/W. Water-cooled long-arc xenon and krypton lamps are made with arc lengths up to 0.3 m (1 ft.), bore diameters up to 10 mm (0.39 in.) and wattages up to 12 kW. Their main application is for laser pumping; krypton arc lamps are especially suitable for pumping neodymium-doped yttrium-aluminum garnet (Nd:YAG) lasers. Forced-air-cooled mercury and halide-doped long-arc lamps are available in lengths up to 1.2 m (4 ft), bore diameters up to 10 mm (0.39 in.), and wattages up to 5 kW. They are used for UV photochemical applications, including the curing of paints, varnishes, and coatings. Mercury capillary lamps are made with arc lengths up to 150 mm (6 in.), bore diameters from 0.08 in. (2 mm), and wattages up to 6 kW. They are used for UV photoexposure in the semiconductor and other industries. They are also finding use in the rapid thermal processing of silicon wafers. All linear arc lamps use special ballasts and high-voltage starting devices. Manufacturers' recommendations for operation should be carefully followed.

ELECTROLUMINESCENT LAMPS85-88 An electroluminescent lamp is a thin (typically less than 1.2 mm [0.05 in.]), flat-area source in which light is produced by a phosphor excited by a pulsating electric field. These lamps are used in decorative lighting, instrument panels, switches, emergency lighting and signs, and for backlighting liquid crystal displays (LCDs). An electroluminescent lamp is a plate capacitor with a phosphor embedded in its dielectric and with one or both of the electrode plates translucent or transparent (Figure 1-20). Green, blue, amber, yellow, or white light is produced by choice of phosphor. The green phosphor has the highest luminance. Lamps are available in rigid ceramic or flexible plastic and are easily fabricated into simple solid or complex multisegmented shapes in lengths as long as 1500 ft. The lamps have thin profiles, light weight, and generate little heat. Some can operate in exterior conditions from −40°C to 121°C (− 40°F to 250°F). The longevity of these lamps depends on isolating the plates and dielectric from humidity during manufacture and use. Their luminance varies with applied voltage, frequency, temperature, and time as well as with the type of phosphor and lamp construction. Electroluminescent lamps can be dimmed. The relationship between voltage and luminance for ceramic and plastic electroluminescent lamps is shown in Figure 6-82. Unlike that of some light sources, the color does not change as the voltage is increased or decreased. At 120 V, 60 Hz, the luminance of the ceramic form with the green phosphor is approximately 0.33 cd/ft2 (3.5 cd/m2); the luminance of the plastic form can be as high 2.5 cd/ft2 (27 cd/m2) under these conditions, or up to 11.6 cd/ft2 (125 cd/m2) at 120 V, 400 Hz. With the ceramic form at 600 V, 400 Hz, a luminance of 6.5 cd/ft2 (70 cd/m2) has been achieved.

Figure 6-82. Luminances of green ceramic and plastic electroluminescent lamps operated at 400 Hz as a function of voltage. Electroluminescent lamps have long life and low power requirements. They usually do not fail abruptly, and because of this, useful life is taken as the number of operating hours after which luminance falls below a specified level. Some electroluminescent lamps using long-life phosphor have operated continuously for more than ten years while powered at 115 V, 400 Hz. The number of operating hours after which the luminance falls to 50% of initial has been used for comparing lamp performance, but this can vary greatly depending on the drive voltage and frequency and the type of phosphor used (Figure 6-83). The value of the time to half luminance for flexible-type lamps operated at 115 V, 400 Hz can vary from 1000 to 30,000 h, depending on the type of phosphor. Typical initial current and power at these parameters are 1.2 mA and 40 mW/in.2

Figure 6-83. Light output versus hours of operation for green ceramic and plastic electroluminescent lamps. The types of electroluminescent devices explained above require high-voltage drivers, which could be a negative feature for use in certain applications. A new type of electroluminescent material known as light emitting polymer (LEP) that operates at low voltage was discovered in 1990. LEPs are organic semiconducting materials that exhibit light emitting characteristics similar to conventional inorganic semiconductors such as light emitting diodes (LEDs) (see section below). In addition, they possess the desirable processing and mechanical characteristics of plastics. A key feature of LEP is that it emits light in a way that is similar to conventional LEDs; in addition, the emitted light can be patterned like LCDs. Therefore, this technology lends itself to the creation of ultrathin lighting displays that operate at low voltage. LEP technology is still in its infancy. Monochromatic devices that are suitable for small display lighting, such as digital readouts on electronic devices, are beginning to emerge commercially.

LIGHT-EMITTING DIODES AlInGaP and InGaN Light-Emitting Diodes89-96 Aluminum indium gallium phosphide (AlInGaP) and indium gallium nitride (InGaN) are the two most common light emitting diode (LED) technologies, displacing older gallium arsenide phosphide (GaAsP), gallium phosphide (GaP), and aluminum gallium arsenide (AlGaAs) LEDs. The following is an overview of these technologies from the perspective of illuminating engineering. Additional information is available for AlInGaP LEDs (References 89-92) and InGaN LEDs (References 93-94), and for LEDs in general (References 95-96). Details on the physics of LEDs is available in Chapter 1, Light and Optics.

Intensity and Color Manufacturers commonly test and bin each LED for luminous intensity and color. Like incandescent reflector lamps, the luminous intensity of LED devices is specified in terms of its beam angle. However, LED manufacturers refer to this as the viewing cone angle or the 2θ—12— angle (in degrees). Typically, all LEDs within a bin do not vary in luminous intensity by more than an factor of two. Because of the difficulty in measuring luminous intensity accurately, there is an expected 10% overlap between adjacent bins. The color of an LED device is specified in terms of the dominant wavelength emitted, λd (in nanometers). Amber AlInGaP and InGaN blue and blue-green LEDs are binned by dominant wavelength, λd. Mixing two color bins in the same pixel matrix can produce an uneven color appearance and is not recommended. Dominant wavelengths for the colors produced by AlInGaP and InGaN LED devices are plotted on the 1931 CIE chromaticity diagram, shown in Figure 6-84. AlInGaP LEDs produce the colors red (626 to 630 nm), red-orange (615 to 621 nm), orange (605 nm), and amber (590 to 592 nm). InGaN LEDs produce the colors green (525 nm), blue green (498 to 505 nm), and blue (470 nm).

Figure 6-84. Dominant wavelengths, λd (nm), plotted on the 1931 CIE Chromaticity Diagram. Typical spectral half-bandwidths are usually specified on LED lamp data sheets as ∆λ1/2. For AlInGaP LEDs, the typical half-bandwidth is approximately 17 nm, and for InGaN LEDs it is approximately 35 nm. The luminous intensity, color, and forward voltage of AlInGaP LEDs are affected by the temperature of the LED p-n junction. As the temperature of the LED pn junction increases, the luminous intensity decreases, the dominant wavelength shifts towards longer wavelengths, and the forward voltage drops. The variation in luminous intensity of InGaN LEDs with operating ambient temperature is small (about 10%) from − 20°C to 80°C. The small variations are not readily visible and need not be taken into account for most applications. The dominant wavelength of InGaN LEDs does vary with LED drive current; as the LED drive current increases, dominant wavelength moves toward shorter wavelengths.

Dimming LEDs may be dimmed to 10% of maximum by reducing the drive current and still have even luminous intensity across an LED matrix. Pulse width modulation (PWM) is the preferred method for dimming LEDs. LEDs may be dimmed to 0.05% of maximum by using PWM. With PWM, the peak pulse current and the pulse rate remains constant while the duration of the on-time pulse is shortened.

Reliability The rated maximum junction temperature (TJMAX) is the most critical parameter on an LED device data sheet. Temperatures exceeding this value usually result in catastrophic failure of a plastic encapsulated LED device. TJMAX is actually a limitation related to packaging rather than to the LED chip. Lamp life for LED devices is based on the mean time between failures (MTBF). MTBF is determined by operating a quantity of LED devices at rated current in an ambient temperature of 55°C and recording when half the devices fail. Lumen depreciation for AlInGaP LED devices is shown in Figure 6-85. The dotted line is an extrapolation from currently available data. The same information for InGaN LED devices is shown in Figure 6-86. Fewer data are available for this technology, and those data were collected at 25°C (77°F). The dotted line is extended for currently available data.

Figure 6-85. Projected long-term light output degradation for AlInGaP LED technology at 55¡C.

Figure 6-86. Projected long-term light output degradation for InGaN LED technology at 25¡C.

AlInGaP and InGaN LED Package Configurations Amber, orange, and red AlInGaP plastic T-1 3/4 LED lamps (Figure 6-87a) with circular radiation patterns have been the preferred package for high-luminance applications such as traffic signals, variable message signs, commercial advertising signs, and EXIT signs. InGaN blue-green plastic T-1 3/4 LED lamps are used in green traffic signals. InGaN blue and green and AlInGaP amber and red plastic T-1 3/4 oval LED lamps with elliptical radiation patterns are used in color graphical commercial LED advertising signs and large scale outdoor full color video displays. Higher intensity AlInGaP amber and red LEDs (Figure 687b) were designed primarily for automotive exterior lighting, such as tail lights and turn signals. Now other high-luminance designs are incorporating these LED devices. New plastic package configurations (Figure 6-87c) incorporating both AlInGaP and InGaN LED devices, have been designed to illuminate large areas. AlInGaP amber and red plastic LED surface mount (SMT) emitters (Figure 6-88a and b) are popular on PC board assemblies. The brightness of chip LED devices is low and useful only for applications where the ambient light level is also low. The subminiature LED devices provide higher brightness and can be used in more applications.

White LED Devices Efforts have been undertaken to develop LED devices that produce white light. One method is to combine red, green, and blue LED chips in the same package to produce white light. However, individuals with color deficiencies may not see the emitted light as white. Recently InGaN LED devices have been combined with photoluminescent phosphors (see Chapter 1, Light and Optics, for information on photoluminescence). The short-wavelength energy from the InGaN LED device induces fluorescence in the phosphor that is encapsulated in the epoxy surrounding the LED chip. The photoluminescent phosphor emits a broad spectral distribution, which in combination with the spectrum of a blue LED, produces a blue-white color (Figure 6-89). Color deficient individuals will still see the combined spectral distribution as blue-white light.

Figure 6-87. Plastic package high-performance LED devices.

Figure 6-86. Projected long-term light output degradation for InGaN LED technology at 25¡C.

Figure 6-88. Plastic package surface-mount (SMT) LED devices.

Figure 6-89. Spectral distribution of a blue-white LED/photoluminescent phosphor device.

Photometric and Electrical Characteristics of LEDs Because they are highly directional light sources, LEDs often are specified in terms of their peak intensity, in millicandelas (mcd). Depending on their viewing cone angle, peak intensities of high-output AlInGaP LEDs range from 100 to 650 mcd for red lamps, and from 120 to 1100 mcd for amber ones. Blue InGaN LEDs with narrow viewing cone angles can have peak intensities greater than 500 mcd. Typical operating voltages for LEDs can range from 1.5 to 4 V. With a typical current of 20 mA, the input power of a single LED can range from approximately 0.03 to 0.08 W. A single red traffic signal head might contain from 20 to 200 individual LED components and have an input power of 10 to 15 W. Luminous efficacy of LEDs is defined as the emitted luminous flux (in lm) divided by the emitted radiant power (in W). This is commonly called internal efficacy. Using this definition, blue LEDs can have a rated internal efficacy in the order of 75 lm/W; red LEDs, approximately 155 lm/W; and amber ones, 500 lm/W. Taking into consideration losses due to internal re-absorption, the luminous efficacy is on the order of 20 to 25 lm/W for amber and green LEDs. This definition of efficacy is called external efficacy and is analogous to the definition of efficacy typically used for other light source types. See Chapter 1, Light and Optics, for definitions of these terms most relevant to illuminating engineering.

LASERS The word "laser" is an acronym for "light amplification by stimulated emission of radiation." Invented in 1960, a laser is a device that concentrates light waves on an intense, low-divergence beam. Even though the light source is an inefficient converter of electrical energy to light energy, a single laser becomes incredibly efficient when applied to a very large-scale lighting requirement (Figure 6-90).

Figure 6-90. Laser display equipment (left, optical components; center, laser head; right, display unit mounted for remote control). What is collectively referred to as a laser is really a complete lighting system comprised of three main parts: the laser tube, a gas-filled tube that emits the light; the projector that manipulates or scans the beam; and the computer hardware and software that stores and controls the performance. Chapter 1, Light and Optics, gives details on laser physics. Lasers are used in hundreds of different applications ranging from corporate theater and major concert tours to surveying and construction. Industrial business theater presentations use sophisticated laser graphics and animation to augment multi-image slide and video productions. Laser effects highlight or emphasize a corporate message or speaker with dramatic flair. Many large corporations use lasers at their conventions and special events. Performers have incorporated lasers in their performances and music videos. Large-scale uses have included theme parks. The laser tube filled with either argon or krypton gas emits the visible light to be manipulated into spectacular light sculptures and paintings. Argon lasers emit light in the blue-green range. For aerial projections to be visible, all the light energy from large-frame 20-watt lasers must be used. For projected images, however, the argon beam can be divided into its component wavelengths of blue and green to create beam colors that can be manipulated separately. Krypton lasers emit red light, which provides important color contrast and spectrum expansion for facade projections. Laser installations and performances are regulated by federal, state, and local regulations designed to protect public health and safety. The concentrated energy in the low-divergence beam (within a determinable distance from the beam origin) could cause retinal damage if projected directly into the eye. Consequently, various regulations call for minimum separation distances (e.g., 3-meter vertical, 2.5-meter lateral) between beam projections and humans.

NUCLEAR LIGHT SOURCES Nuclear light sources are self-contained and require no power supply. They typically provide illumination for instrument panels, controls, clocks, and exit signs (see Chapter 29, Emergency, Safety, and Security Lighting). These sources consist of a sealed glass tube or bulb internally coated with a phosphor and filled with tritium gas. Low-energy beta particles (electrons) from the tritium, an isotope of hydrogen, strike the phosphor, which in turn emits light of a color characteristic of the type of phosphor used. Thus, the mechanism of light production is very similar to that in a conventional television tube. The higher the ratio of the quantity of tritium to the phosphor area, the greater the luminance. The luminance can range up to 7 cd/m2 (0.65 cd/ft2), with a typical average of 1.7 cd/m2 (0.16 cd/ft2) (this level is approximately that of an illuminated car instrument panel), and the sources can be supplied in a variety of colors. Highest luminances are obtained in the green and yellow phosphors, and green is the usual color supplied. Since tritium has a half-life of 12.3 yr, one might expect the brightness of the source to decay likewise. In reality, the half luminance typically is reached in 6 to 7 yr, and the useful life of these lamps is currently about 15 yr. These light sources can be supplied in very small sizes, down to 5 mm (0.2 in.) in diameter by 2 mm (0.079 in.) in length. The glass wall is impervious to tritium and completely absorbs any beta radiation not already absorbed by the phosphor. The unit is thus a completely sealed source and does not present any radiation hazard. Glass capsules can be produced in a wide variety of shapes and sizes and are usually made to normal glassworking tolerances. All applications of these lamps are monitored by the Nuclear Regulatory Commission in the United States and by Atomic Energy Canada in Canada.

CARBON ARC LAMPS Carbon arc lamps were the first commercially practical electric light sources. They were used for many years in applications where extremely high luminance, high correlated color temperature, and/or high color rendering were necessary, such as in motion picture projection lamphouses and for theatrical followspots, searchlights, and film production daylight supplemental lighting. In most of those applications, xenon short arc and metal halide light sources have replaced carbon arc. Carbon arcs are operated in lamphouses that shield the outside from stray radiation. These lamphouses incorporate optical components such as lenses, reflectors, and filters for eliminating undesired parts of the spectrum. For more information about different types of carbon arc lamps, their operating characteristics, and power sources, see previous editions of the IESNA Lighting Handbook.

GASLIGHTS

Gaslights use gaseous fuels for light and decorative purposes. They use open gas flames or incandescent mantles of the upright and inverted types. For more information, see previous editions of the IESNA Lighting Handbook.

REFERENCES 1. Elenbaas, W. 1972. Light sources. New York: Crane, Russak & Co. 2. Hammer, W. J. 1913. The William J. Hammer historical collection of incandescent electric lamps. New York Electrical Society, New Series no. 4. 3. Schroeder, H. 1923. History of electric light. Smithsonian Miscellaneous Collections 76(2). 4. Coolidge, W. D. 1910. Ductile tungsten. Trans. Am. Inst. Elec. Eng. 24:961-965. 5. Forsythe, W. E., and E. Q. Adams. 1937. The tungsten filament incandescent lamp. J. Sci. Lab., Denison Univ. 6. Smithells, C. J. 1953. Tungsten: Its metallurgy, properties, and applications. New York: Chemical Publishing Co. 7. Rieck, G. D. 1967. Tungsten and its compounds. New York: Pergamon Press. 8. Langmuir, I. 1912. Convection and conduction of heat in gases. Phys. Rev. 34(6):401-22. 9. Thouret, W. E., H. A. Anderson, and R. Kaufman. 1970. Krypton-filled large incandescent lamps. Illum. Eng. 65(4):231-40. 10. Thouret, W. E., R. Kaufman, and J. W. Orlando. 1975. Energy and cost saving krypton filled incandescent lamps. J. Illum. Eng. Soc. 4(4):188-97. 11. Morris, R. W. 1947. Considerations affecting the design of flashing signal filament lamps. Illum. Eng. 42(6):625-35. 12. American National Standards Institute. 1996. Lamp-base temperature rise-Method of measurement, C78.25-1996. New York: ANSI. 13. Forsythe, W. E., E. Q. Adams, and P. D. Cargill. 1939. Some factors affecting the operation of incandescent lamps. J. Sci. Lab., Denison Univ. 14. Merrill, G. S. 1931. Voltage and incandescent electric lighting. Proceedings of the International Illumination Congress, edited by W. S. Stiles, London: International Illumination Congress. 15. Industry Committee on Interior Wiring Design. 1941. Handbook of interior wiring design. New York: Industry Committee on Interior Wiring Design. 16. Potter, W. M., and K. M. Reid. 1959. Incandescent lamp design life for residential lighting. Illum. Eng. 54(12):751-57. 17. Questions and answers on light sources. 1968. Illum. Eng. 63(6):339. 18. Whittaker, J. D. 1933. Applications of silver processed incandescent lamps with technical data. Trans. Illum. Eng. Soc. 28(5):418-36. 19. Evans, M. W., LaGiusa F. F., and J. M. Putz. 1977. An evaluation of a new ellipsoidal incandescent reflector lamp. Light. Des. Appl. 7(3):22-25. 20. Leighton, L. G. 1962. Characteristics of ribbon filament lamps. Illum. Eng. 57(3):121-26. 21. American National Standards Institute. 1998. Miniature and sealed-beam incandescent lamps--Method of designation, C78.390-1998. New York: ANSI. 22. American National Standards Institute. 1997. American standard method for the designation of glow lamps, C78.381-Rev-1997. New York: ANSI. 23. Waymouth, J. F. 1971. Electric discharge lamps. Cambridge: M.I.T. Press. 24. Townsend, M. A. 1942. Electronics of the fluorescent lamp. Trans. Am. Inst. Elec. Eng. 61(8):607-12. 25. Haft, H. H., and W. A. Thornton. 1972. High performance fluorescent lamps. J. Illum. Eng. Soc. 2(10):29-35. 26. Verstegen, J. M. P. J., D. Radielovic, and L. E. Vrenken. 1975. A new generation deluxe fluorescent lamp. J. Illum. Eng. Soc. 4(1):90-105. 27. Denneman, J. W., J. J. de Groot, A. G. Jack, and F. A. S. Ligthart. 1980. Insights into the 26 mm diameter fluorescent lamp. J. Illum. Eng. Soc. 10(1):2-7. 28. Bessone, C. S., and R. J. Citino. 1981. Optimum system and lamp parameters for efficient T8 fluorescent systems. J. Illum. Eng. Soc. 11(1):2-6. 29. Lowry, E. F., W. C. Gungle, and C. W. Jerome. 1954. Some problems involved in the design of fluorescent lamps. Illum. Eng. 49(11):545-52. 30. Lowry, E. F., W. S. Frohock, and G. A. Meyers. 1946. Some fluorescent lamp parameters and their effect on lamp performance. Illum. Eng. 41(12):859-71. 31. Lowry, E. F. 1952. The physical basis for some aspects of fluorescent lamp behavior. Illum. Eng. 47(12):639-46. 32. Lowry, E. F., and E. L. Mager. 1949. Some factors affecting the life and lumen maintenance of fluorescent lamps. Illum. Eng. 44(2):98-105. 33. Ouellette, M., B. Collins, and S. Treado. 1993. The effect of temperature on starting and stabilization of compact fluorescent systems. Conference Record of the IEEE Industry Applications Society 28th Annual Meeting, Piscataway, NJ: Institute of Electrical and Electronics Engineers. 34. McFarland, R. H., and T. C. Sargent. 1950. Humidity effect on instant starting of fluorescent lamps. Illum. Eng. 45(7):423-28. 35. Hammer, E. E. 1981. Peak and RMS starting voltage procedure for standard/low energy fluorescent lamps. J. Illum. Eng. Soc. 10(4):204-10. 36. Hammer, E. E. 1983. Fluorescent lamp starting voltage relationships at 60 HZ and high frequency. J. Illum. Eng. Soc. 13(1):36-46. 37. Hammer, E. E. 1991. Fluorescent system interactions with electronic ballasts. J. Illum. Eng. Soc. 20(1):56-63. 38. Hammer, E. E. 1984. Fluorescent lamp operating characteristics at high frequency. J. Illum. Eng. Soc. 14(1):211-24.

39. Aoike, N., K. Yuhara, and Y. Nobuhara. 1984. Electronic ballast for fluorescent lamp lighting system of 100 lm/W overall efficiency. J. Illum. Eng. Soc. 14 (1):225-39. 40. American National Standards Institute. 1995. Fluorescent lamp ballasts--Methods of measurement, ANSI C82.2-1995. New York: ANSI. 41. He, Y. 1998. National Lighting Product Information Program. Specifier reports: Lighting Circuit Power Reducers. Troy, NY: Lighting Research Center, Rensselaer Polytechnic Institute. 42. American National Standards Institute. 1993. High-frequency Fluorescent lamp ballasts, ANSI C82.11-1993. New York: ANSI. 43. Carpenter, W. P. 1951. Application data for proper dimming of cold cathode fluorescent tubing. Illum. Eng. 46(6):306-9. 44. Campbell, J. H., H. E. Schultz, and W. H. Abbott. 1954. Dimming hot cathode fluorescent lamps. Illum. Eng. 49(1):7-14. 45. Von Zastrow, E. E. 1963. Fluorescent lamp dimming with semiconductors. Illum. Eng. 58(4):312-17. 46. Campbell, J. H., and D. C. Kershaw. 1956. Flashing characteristics of fluorescent lamps. Illum. Eng. 51(11):755-60. 47. Bunner, R. W., and R. T. Dorsey. 1956. Flashing applications of fluorescent lamps. Illum. Eng. 51(11):761-67. 48. Elenbaas, W. 1951. The high pressure mercury discharge. New York: Interscience. 49. Elenbaas, W. 1965. High pressure mercury-vapor lamps and their applications. [Eindhoven]: Philips Technical Library. 50. Till, W. S., and M. Pisciotta. 1959. New designations for mercury lamps. Illum. Eng. 54(9):594-96. 51. Till, W. S., and M. C. Unglert. 1960. New designs for mercury lamps increase their usefulness. Illum. Eng. 55(5):269-81. 52. Jerome, C. W. 1961. Color of high pressure mercury lamps. Illum. Eng. 56(3):209-14. 53. Larson, D. A., H. D. Fraser, W. V. Cushing, and M. C. Unglert. 1963. Higher efficiency light source through use of additives to mercury discharge. Illum. Eng. 58(6):434-39. 54. Martt, E. C., L. J. Smialek, and A. C. Green. 1964. Iodides in mercury arcs: For improved color and efficacy. Illum. Eng. 59(1):34-42. 55. Waymouth, J. F., W. C. Gungle, J. M. Harris, and F. Koury. 1965. A new metal halide arc lamp. Illum. Eng. 60(2):85-88. 56. Reiling, G. H. 1964. Characteristics of mercury vapor-metallic iodide arc lamps. J. Opt. Soc. Am. 54(4):532-40. 57. Kühl, B. 1964. High pressure mercury lamps with iodide additives. Lichttechnik 16(2):68-71. 58. Fromm, O. C., J. Seehawer, and W. J. Wagner. 1979. A metal halide high pressure discharge lamp with warm white colour and high efficacy. Light. Res. Tech. 11(1):1-8. 59. Waymouth, J. F. 1971. Electric discharge lamps. Cambridge: M.I.T. Press. 60. Barosi, A., and E. Rabusin. 1974. Zirconium-aluminum alloy as a getter for high intensity discharge lamps. Japanese Journal of Applied Physics Supplement 2 (Part 1):49-52. 61. Gungle, W. A. and, J. F. Waymouth. 1970. High pressure electric discharge device with barium peroxide getter and getter mounting structure, U.S. Patent no. 3,519,864. In Official Gazette of the U. S. Patent Office. 876(1):291. 62. U.S. Food and Drug Administration. [Latest issue]. Performance standards for light-emitting products: Mercury vapor discharge lamps, 21 CFR 1040.30. Washington: U. S. GPO. 63. Keeffe, W. M., Z. K. Krasko, J. C. Morris, and P. J. White. 1988. Improved low wattage metal halide lamp. J. Illum. Eng. Soc. 17(2):39-43. 64. Krasko, Z. K., and W. M. Keefe. 1990. A new M100 metal halide lamp with improved color rendering properties. J. Illum. Eng. Soc. 19(1):118-24. 65. Fromm, D. C., and J. Heider. 1991. Color rendering, color shift, and lumen maintenance of low-wattage metal halide lamps. J. Illum. Eng. Soc. 20(1):77-83. 66. American National Standards Institute. 1990. Specifications for 400-watt S51 high-pressure sodium lamps, ANSI C78-1350-1990. New York: ANSI. 67. IESNA. 1995. Approved method for life testing of high intensity discharge (HID) lamps, LM-47-1995. New York: Illuminating Engineering Society of North America. 68. American National Standards Institute. 1996. Reference ballasts for high intensity discharge lamps--Methods of measurement. ANSI C82.6-1996. New York: ANSI. 69. Rompe, R., and W. E. Thouret. 1938. Quecksilberdampflampen hoher Leuchtdichte [Mercury vapor lamps of high brightness]. Zeitschr. Tech. Physik 19 (11):352-55. 70. Rompe, R., W. E. Thouret, and W. Weizel. 1944. Zur Frage der Stabilisierung frei brennender Lichtbögen [The problem of stabilisation of free burning arcs]. Zeitschr. Physik 122:1-24. 71. Schulz, P. 1947. Elektrische Entladungen in Edelgasen bei hohen Druken [Xenon short arc lamps]. Ann. Physik 1(1-3):95-118. 72. Bourne, H. K. 1948. Discharge lamps for photography and projection. London: Chapman & Hall. 73. Thouret, W. E. 1950. New designs of quartz lamps. Lichttechnik 2:73. 74. Thouret, W. E., and G. W. Gerung. 1954. Xenon short arc lamps and their application. Illum. Eng. 49(11):520-526.

75. Anderson, W. T. 1954. High brightness xenon compact arc lamps. J. Soc. Mot. Pict. Tel. Eng. 63(3):96-97. 76. Thouret, W. E. 1960. Tensile and thermal stresses in the envelope of high brightness high pressure discharge lamps. Illum. Eng. 55(5):295-306. 77. Retzer, T. C. 1958. Circuits for short-arc lamps. Illum. Eng. 53(11):606-12. 78. Thouret, W. E., and H. S. Strauss. 1962. New designs demonstrate versatility of xenon high-pressure lamps. Illum. Eng. 57(3):150-158. 79. Lienhard, O. E., and J. A. McInally. 1962. New compact-arc lamps of high power and high brightness. Illum. Eng. 57(3):173-76. 80. Thouret, W. E., H. S. Strauss, S. F. Cortorillo, and H. Kee. 1965. High brightness xenon lamps with liquid-cooled electrodes. Illum. Eng. 60(10):339-47. 81. Lienhard, O. E. 1965. Xenon compact-arc lamps with liquid-cooled electrodes. Illum. Eng. 60(5):348-52. 82. Thouret, W. E., J. Leyden, H. S. Strauss, G. Shaffer, and H. Kee. 1972. Twenty to 30 kW xenon compact arc lamps for searchlights and solar simulators. J. Illum. Eng. Soc. 2(10):8-18. 83. Lemons, T. M. 1978. HMI lamps. Light. Des. Appl. 8(8):32-37. 84. Hall, R., and B. Preston. 1981. High-power single-ended discharge lamps for film lighting. J. Soc. Mot. Pict. Tel. Eng. 90(8):678-85. 85. Payne, E. C., E. L. Mager, and C. W. Jerome. 1950. Electroluminescence: A new method of producing lighting. Illum. Eng. 45(11):688-93. 86. Ivey, H. F. 1960. Problems and progress in electroluminescent lamps. Illum. Eng. 55(1):13-23. 87. Blazek, R. J. 1962. High brightness electroluminescent lamps of improved maintenance. Illum. Eng. 57(11):726-29. 88. Weber, K. H. 1964. Electroluminescence: An appraisal of its short-term potential. Illum. Eng. 59(5):329-36. 89. Craford, G., and F. Steranka. 1994. Light emitting diodes. In Encyclopedia of applied physics, edited by George L. Trigg. Vol. 8. VCH Publishers. 90. Huang, K. H., J. G. Yu, C. P. Kuo, R. M. Fletcher, T. D. Osentowski, L. J. Stinson, and M. G. Craford. 1992. Twofold efficiency improvement in high performance AlGaInP light-emitting diodes in the 555-620 nm spectral region using a thick GaP window layer. Applied Physics Letters 61(9):1045-47. 91. Craford, M. G. 1992. LEDs challenge the incandescents. IEEE Circuits and Devices 8(5):24-29. 92. Kuo, C. P., R. M. Fletcher, T. D. Osentowski, M. C. Lardizabal, M. G. Craford, and V. M. Robbins. 1990. High performance AlGaInP visible light-emitting diodes. Applied Physics Letters 57(27):2937-39. 93. Nakamura, S., and G. Fasol. 1997. The blue laser diode, GaN based light emitters and lasers. New York: Springer Verlag. 94. Strite, S., and Morkoc, H. 1992. GaN, AIN and InN: A review. Journal of Vacuum Science & Technology B 10(4):1237-1266. 95. Stringfellow, G. B., and G. Craford. 1997. High brightness light emitting diodes, semiconductors and semimetals. New York: Academic Press. 96. Hewlett-Packard Optoelectronics Applications Staff. 1981. Fiber-optics applications manual. New York: McGraw-Hill.

7 Luminaires A luminaire is a device to produce, control, and distribute light. It is a complete lighting unit consisting of the following components: one or more lamps, optical devices designed to distribute the light, sockets to position and protect the lamps and to connect the lamps to a supply of electric power, and the mechanical components required to support or attach the luminaire. This chapter provides information for both specifiers and manufacturers of luminaires. It describes most common types of luminaires, how they are used, and how their performance is evaluated, and gives a general classification system useful for understanding their application. With the exception of lamps (light sources), the characteristics, design, and manufacture of luminaire components are described. Detailed information for the specific applications of luminaires can be found in the appropriate application chapters.

GENERAL DESCRIPTION Light Sources Luminaires are designed and manufactured for all common types of electric lamps. Luminaires are commonly available for these lamps:

Incandescent filament including tungsten halogen and infrared (heating) lamps Fluorescent Compact fluorescent Induction or electrodeless lamps, including fluorescent and sulfur lamps High-intensity discharge lamps, including metal halide, high-pressure sodium, and mercury Low-pressure sodium lamps

Luminaires are less common for xenon arc and carbon arc lamps. The size, materials, thermal properties, photometric performance, and power requirements of a luminaire depend on the type of lamp used. For example, lamps that produce a large amount of infrared (IR) radiation (heat) require luminaires that are vented for convection, and fluorescent lamps that are sensitive to environmental temperature must be protected from low air temperatures.

Light Control Components The lamps used in some luminaires have integrated light control components. These are usually incandescent and tungsten-halogen lamps with a reflective coating and/or refracting prisms on the bulb. These integral lamp components produce useful beams and patterns of light without any auxiliary optical control. In these cases, most of the light control is provided by the lamp; the luminaire is simply an appliance to hold the lamp, deliver electric power, and perhaps permit the lamp to be aimed in different directions. Most lamps emit light in virtually all directions, and their efficient application requires light control components to collect and distribute the light. Four types of light control components are commonly used: reflectors, refractors, diffusers, and louvers or shields. See the sections "Optical Control" in Chapter 1, Light and Optics, and "Materials Used in Luminaires" in this chapter for detailed discussion of optical control by reflection and refraction. Reflectors. A reflector is a device, usually of coated metal or plastic, that has a high reflectance and is shaped to redirect by reflection the light emitted by the lamp. The surface finish of luminaire reflectors usually is classified as specular, semi-specular, spread, or diffuse. For more information, see "Optical Control" in Chapter 1, Lighting and Optics. Some applications require the reflector to control the light very precisely, so specular or semi-specular reflecting material is used. Metal reflectors are formed and then polished or chemically coated to produce a specular finish. In some cases, metal reflectors are manufactured from metal stock that has already been treated to produce a specular finish. Plastic reflectors are molded and then coated with aluminum by vaporization. Examples of specular reflectors are those used to control the light from a metal halide lamp to produce a narrow beam of light for sports lighting, and the parabolic louvers in fluorescent lamp troffers.

Figure 7-1. Examples of reflectors: (a) linear faceted, coated steel reflector in a strip fluorescent lamp luminaire, (b) and (c) spun specular and grooved aluminum reflectors for a compact fluorescent downlight luminaire, (d) faceted reflector for a floodlight luminaire, and (e) reflector with "kicker" to direct light for wall-wash luminaire. In some luminaires the reflector does not have to control the light very precisely, and it is sufficient for the reflector to have a high but nondirectional reflectance. An example of this is the white, slightly specular, coated metal reflectors in some large fluorescent lamp luminaires. On the other hand, diffuse reflectors have very little effect on the distribution of light and are uncommon in luminaires. Other applications and lamps require reflectors with special surface finishes, such as semi-specular or peened materials, or coatings to reduce color separation upon reflection (iridescence) when using certain fluorescent lamps. Examples of reflectors are shown in Figure 7-1. In some cases, reflectors have properties varying with wavelength. Alternating layers of materials with differing indices of refraction are applied to glass. These layers have a thickness approximately that of the wavelength of light (500 nm). Interference effects produce reflection that changes with wavelength. This is useful if it desirable to reflect light but not reflect long-wavelength thermal radiation or, conversely, to reflect the long wavelength radiation and pass light. These reflectors are used when it is necessary to direct light and control the heat generated by the lamps. Refractors. Refractors are light control devices that take advantage of the change in direction that light undergoes as it passes through the boundary of materials of differing optical density (index of refraction), such as air to glass or air to plastic (see Figure 7-2). A material, usually glass or plastic, is shaped so that light is redirected as it passes through it. This redirecting can be accomplished with linear (extruded, two-dimensional) prisms or with three-dimensional

pyramidal-shaped prisms. These prisms can be either raised from the surface of the material or embossed into it. They are usually small enough to become a type of surface treatment on one side of an otherwise flat sheet of glass or plastic. The entire sheet is referred to as a prismatic lens.

Figure 7-2. Examples of refractors: (a) prismatic lens on surface-mounted fluorescent lamp luminaire, (b) recessed luminaire with spread lens, (c) glass refractor on an outdoor area luminaire, (d) Fresnel refractor, (e) wraparound prismatic lens on a fluorescent lamp luminaire, (f) prismatic lens on recessed flourescent lamp luminaire, (g) low-bay industrial luminaire with prismatic refractor, and (h) track luminaire with spread lens refractor. A collection of small prisms, acting in concert, can be used to control the directions from which light leaves a luminaire. This redirection can be used to partially destroy images and therefore to obscure lamps and reduce luminance by increasing the area over which the light leaves the luminaire. In some cases the sheet containing prisms is shaped to provide additional control. In specialized applications, such as the refractors used for some street lighting luminaires, the prisms are on both surfaces of the material. Another application of refracting material takes advantage of total internal reflection. In this case the refracting material is shaped so that light passes into it through its first surface and mostly is reflected from the second surface back into the material and out the first surface. Some glass and plastic industrial luminaires use this type of light control. This is also the basis for the operation of light pipes and fiber-optic luminaires. For some luminaires, the lamp and application require a transparent cover to block ultraviolet (UV) radiation or prevent broken lamp components from falling out of the luminaire. Though providing little optical control, these cover plates often are referred to as lenses. Diffusers. Diffusers are light control elements that scatter (redirect) incident light in many directions. This scattering can take place in the material, such as in bulk diffusers like white plastic, or on the surface as in etched or sandblasted glass. Diffusers are used to spread light and, since scattering destroys optical images, obscure the interior of luminaires, suppress lamp images, and reduce high luminances by increasing the area over which light leaves a luminaire. Examples of diffusers are shown in Figure 7-3. Shades, Shields, Louvers, and Baffles. Shades and shields are opaque or translucent materials shaped to reduce or eliminate the direct view of the lamp from outside the luminaire (Figure 7-4). Shades are usually translucent and are designed to diffuse the light from the lamp and provide some directional control. Blades, usually opaque and highly reflective, can be shaped and positioned to eliminate the direct view of the lamp from certain directions outside the luminaire and to control the direction from which the light leaves.

Figure 7-3. Examples of diffusers: (a) and (b) wrap-around white diffusers for fluorescent lamp luminaires, (c) jelly jar diffuser for compact fluorescent lamp luminaire, and (d) drop glass diffuser for metal halide lamp luminaire.

Figure 7-4. Examples of baffles, louvers, and shields: (a), (b), and (c) louvers for fluorescent lamp luminaires; (d) cross baffles for compact fluorescent lamp luminaires; (e) shield for industrial fluorescent lamp luminaire; and (f) hoods and cowls for track luminaires. If arranged in a rectangular grid, producing cells, they are called louvers. If arranged linearly they are called baffles. In large fluorescent lamp luminaires, louvers are designed so that the lamps are directly above the center of the cells formed by the louvers. In long narrow fluorescent lamp luminaires, baffles extend across the axis of the lamps. Louvers and baffles often are made of specularly reflecting metal, though some are of coated plastic. Though intended to eliminate the direct view of the lamp at some angles, specular louvers and baffles can provide lamp images at other viewing angles by reflection. In turn, these images may produce images in the

screen of computer VDTs. See Figure 7-4 for examples of baffles and louvers. Fiber-Optic Luminaires. A fiber-optic illumination system is a distributed lighting system allowing remote source illumination of areas and objects. A fiberoptic lighting system has a light source or illuminator, optical fiber, and various output fixtures selected to illuminate specific areas or objects. These systems can be used in lighting applications requiring the light source and light output to be separated, as in hazardous environments, wet locations, or in temperaturesensitive spaces. They also can be used when it is desirable to have sizes, shapes, or light output characteristics different from conventional luminaires. Virtually all of the IR and most of the UV radiation from lamps, as well as the electrical connections needed to power the system, are absent from the illuminated space. The separation of source from output allows the use of various components within the source enclosure that can provide interesting optical effects at the output. Such components can include color or effects wheels and filters. Illuminator. The illuminator is an assembly that houses the light source and positions the optical fibers within an output port, or ports, with respect to the source. It may include other components such as a cooling fan, various optical elements, filters, and color wheels or effects devices. Since lamps are not point sources, the light is collected with a combination of reflecting and/or refracting optics. The collection efficiency is a function of both the size and shape of the light source (e.g., arc or filament) and the chemical as well as the physical characteristics of the fiber(s) receiving the light. Proper use of conventional nonimaging optical design methods enable the collection of 10 to 40% of the light into light guides. The most commonly used lamps in optical fiber systems are tungsten-halogen and metal halide. Illuminators that incorporate halogen sources almost always includes a cooling fan. Applications that require greater light output and longer lamp life use metal halide high lamps. Other light sources (xenon, xenon metal halide, and sulfur lamps) occasionally are used in optical fiber illumination systems. These lamps have advantages that make them appropriate for certain specialized applications. The illuminator often contains various accessories that affect the output. These include permanent filters such as IR, UV, or dichroic; rotating wheels that provide a variety of colors; and mechanical dimming or twinkling. Many tungsten-halogen illuminators are available with electronic dimming. Optical Fiber. The principal function of optical fiber is to deliver light from the illuminator to the exit port. A fiber consists of at least two and often three concentric regions. These regions are the core, which is the light transmission medium; the cladding, which confines the light to the core, and the sheathing, which is an outer coating protecting the fiber from handling or interaction with the environment. In side lighting applications, the fiber is designed so that light is emitted uniformly from the sides. For end lighting applications, light propagates down the fiber by totally internally reflecting at the core-cladding interface. Since this reflectance is very nearly 1.0, light loss through the fiber is due almost solely to absorption and scattering. The absorption and scattering characteristics of the fiber material determine the attenuation. Total internal reflection requires a high incidence angle to the core-cladding interface, and this requires that light enter the fiber end within a certain acceptance angle. The acceptance angle is a function of the indices of refraction of the core and cladding. In general, the acceptance angle also defines the angle with which the light exits an optical fiber. Output Optics. The nature of the optics placed on the output end of the fibers depends on the details of the application. The angular distribution of the light depends on the spread of the system's input illumination, and is typically between 60° and 80°. A wide array of fixtures that house shaped plastic or glass diffusers and/or lenses with different distribution characteristics is available to achieve many effects. In addition, numerous decorative fiber end fittings are available to create interesting display and decorative effects. Optical fiber output fixtures typically are smaller and more lightweight than conventional luminaires because the fibers approach the characteristics of a directional point source and heat is of no concern. This allows many fixtures to be composed of plastic.

Mechanical Components The mechanical components of a luminaire consist of a housing or general structure to support other components of the luminaire, and a mounting mechanism for the attachment of the luminaire to its support (Figure 7-5). In some luminaires the reflector is a separate component that is attached to the housing, as in a compact fluorescent lamp downlight. In other luminaires, the housing serves as the reflector, as in a fluorescent lamp troffer.

Figure 7-5. Examples of mechanical components of luminaires: (a) and (b) fluorescent lamp troffers showing housing and mounting to inverted-T ceiling system; (c) compact fluorescent lamp downlight showing housing, mounting for ballast, and mounting brackets; and (d) mounting and electrical connection for a pendant-mounted luminaire. If the luminaire uses a refractor or transparent cover, then hinged frames or doors often are provided to hold the lens. Access for cleaning and relamping is through this door. In damp or wet applications it is necessary to provide adequate seals to prevent migration of water into the luminaire. In some hazardous locations the housing and seals must keep explosive or flammable vapors from contact with high lamp surface temperatures or electric spark. These luminaires are said to be

explosion proof. Many recessed luminaires are vented to dissipate heat that can degrade lamp performance. In some applications, the luminaire is used as part of the building's heating, ventilating, and air conditioning system. Air is supplied to or removed from the room using the luminaire. In this case, airways are provided within the luminaire as well as attachments for air ducts and slots through which air enters or leaves the room.

Electrical Components The electrical components of the luminaire operate the lamp (Figure 7-6). One or more sockets provide mechanical support for the lamp and furnish necessary electrical connections. For some lamps, usually single-ended, mechanical support is required in addition to the socket.

Figure 7-6. Electrical components, showing junction box, ballast, and lamp socket for (a) metal halide and (b) compact fluorescent lamp luminaires; (c) compact fluorescent lamp, socket, magnetic ballast, and connectors; and (d) photocell, transformer, and ballast for controlling outdoor luminaires. If required, the luminaire contains and supports ballasts, starters, igniters, capacitors, or emergency lighting devices. The size and power handled by these components often determine the size of the luminaire and the requirements for proper thermal performance. In a few applications, these components are too heavy, too loud, or too large to be in the luminaire. In these cases, the ballast and other auxiliary equipment are mounted remotely from the luminaire and lamp. The luminaire also contains wiring and connectors to connect the lamp socket and, if present, the ballast to the external wiring that brings electrical power to the luminaire. Figures 7-7 through 7-12 show cross sections of typical luminaires with most of the major components shown. Even apparently simple luminaires contain many components.

Figure 7-7. Incandescent lamp downlight showing housing, mounting, reflector, wiring, socket, and lamp.

Figure 7-8. Compact fluorescent lamp downlight showing housing, mounting, reflector, wiring, socket, and lamp.

Figure 7-9. Fluorescent lamp troffer showing housing, mounting, reflector, lamps, and ballast.

Figure 7-10. Suspended fluorescent lamp luminaire showing extruded aluminum housing, reflector, lamps, and ballast.

Figure 7-11. High-bay HID lamp luminaire showing housing, reflector, lamp, socket, magnetic ballast and capacitor, and mounting.

Figure 7-12. Fiber-optic luminaires.

LUMINAIRE TYPES AND CLASSIFICATION Purpose of Classification Luminaire classification helps specifiers and manufacturers describe, organize, catalog, and retrieve luminaire information. The nature of luminaire classification has changed with the advance of computer and information technology. Modern lighting design and specification practice relies on computerbased luminaire databases, accessed on CD-ROM or over the Internet. This technology allows luminaire data to be updated frequently and easily. In such systems, a luminaire can be known by all of its characteristics, with any one being the path by which a search finds the luminaire in a database.

Methods for Classification Luminaires can be classified according to source, mounting, construction, application, and/or photometric characteristics. Classifications by application and photometric characteristics are discussed in the next two sections. Classification by Application. A common form of classification organizes luminaires by application. Many luminaire characteristics are determined by application, so this distinction proves useful in organizing luminaire information. Three application areas are usually distinguished: residential, commercial, and industrial. Within each application, luminaires can be classified by source, mounting, and con struction. Examples of these include residential ceiling-mounted room luminaires using incandescent lamps, recessed fluorescent lamp troffer luminaires, and high bay suspended metal halide lamp luminaires. Classification by Photometric Characteristics. Another form of classification uses the luminous intensity or flux distribution of the luminaire. For luminaires used indoors, a method specified by the International Commission on Illumination (CIE) is commonly used. For luminaires used outdoors, the NEMA and

IESNA methods are used. The CIE Classification System. The International Commission on Illumination provides a classification system based on the proportion of upward and downward directed light output. This system is usually applied to indoor luminaires.

Direct lighting. When luminaires direct 90 to 100% of their output downward, they form a direct lighting system. The distribution may vary from widespread to highly concentrated, depending on the reflector material, finish, and contour and on the shielding or optical control media employed. Semidirect lighting. The distribution from semidirect units is predominantly downward (60 to 90%) but with a small upward component to illuminate the ceiling and upper walls. General diffuse lighting. When the downward and upward components of light from luminaires are about equal (each 40 to 60% of total luminaire output), the system is classified as general diffuse. Direct-indirect is a special (non-CIE) category within this classification, in which the luminaires emit very little light at angles near the horizontal. Semi-indirect lighting. Lighting systems that emit 60 to 90% of their output upward are classified as semi-indirect. Indirect lighting. Lighting systems classified as indirect are those that direct 90 to 100% of the light upward to the ceiling and upper side walls.

Indoor Luminaire Classifications By Cutoff. There are several characteristics of indoor luminaire intensity distributions that are important for classification. This information can appear in the photometric report for a luminaire. See the section "Luminaire Photometric Report" later in this chapter.

Physical cutoff. The angle measured from nadir at which the lamp is fully occluded. Optical cutoff. The angle measured from nadir at which the reflection of the lamp in the reflector is fully occluded. Shielding angle. The angle measured from the horizontal at which the lamp is just visible.

The NEMA Classification System. This system is based on the distribution of flux within the beam produced by the luminaire. It is used primarily for sports lighting and floodlighting luminaires. Seven distributions are defined, types 1 through 7, from narrowest to widest beams (see Figure 20-10 in Chapter 20, Sports and Recreational Area Lighting). The IESNA Classification System For Outdoor Luminaires. This system is based on the shape of the area that is primarily illuminated by the luminaire. It is used for roadway and area lighting luminaires. Though these luminaires can differ in the manner in which they are mounted, by the type of intensity distribution they exhibit, and by the degree to which they provide cutoff, these luminaires often are specified by the way in which they illuminate an area. Following are the IESNA outdoor luminaire classifications by intensity distribution (see Figure 22-6 in Chapter 22, Roadway Lighting). More detailed information on these luminaire types is found in Chapter 22, Roadway Lighting.

Principal Types of Luminaires Commercial and Residential Portable Luminaires. These are completely self-contained luminaires designed to be moved and placed near the task to be lighted. They have a plug and outlet connection to electric power and usually contain integral switching and/or dimming. They usually contain low-wattage incandescent, tungsten-halogen, or compact fluorescent lamps. Examples of portable luminaires are floor and table luminaires, desk luminaires, and partition-mounted luminaires (Figure 7-13).

Figure 7-13. Examples of portable luminaires using (a) compact fluorescent lamp, (b) incandescent lamp, and (c) straight fluorescent lamps.

Figure 7-14. Examples of furniture-mounted luminaires: (a) under-cabinet luminaire with wrap-around prismatic lens, (b) under cabinet luminaire with extruded lens, and (c) partition-mounted luminaire with metal halide lamp.

Figure 7-15. Examples of downlight luminaires using (a) compact fluorescent, (b) incandescent, and (c) metal halide lamps; (d) fluorescent lamp recessed troffer with parabolic louvers; (e) continuous linear; and (f) fluorescent lamp recessed troffer with prismatic lens.

Figure 7-16. Examples of ceiling surface mounted luminaires; (a) wrap-around lens, (b) incandescent lamp downlight, (c) troffer, (d) metal halide lamp area light, and (e) metal halide lamp downlight. Furniture Mounted. Permanently attached to furniture or other equipment surface, these luminaires are designed to be in close proximity of the task and produce localized lighting. They can be found under kitchen cabinets and in bathroom vanities (Figure 7-14). Recessed Downlights. These are general-purpose luminaires designed to provide general or ambient lighting in a space. They are recessed into the ceiling and are designed to produce illuminance on a floor or workplane. Certain types have concentrated luminous intensity distributions designed for spaces with computer VDTs. It is often necessary to augment these luminaires with other types that raise wall luminances and add vertical illuminance to the space. Recessed downlights can be grouped by size. There are two types of recessed downlights (Figure 7-15). Incandescent, compact fluorescent, and metal-halide lamp downlights usually have modest apertures and can exhibit very low luminances at high viewing angles. Fluorescent lamp troffers use large fluorescent lamps and are usually used with a suspended tile ceiling system. Sizes range from 6 in. × 48 in. to 48 in. square. Ceiling Surface Mounted. These luminaires can provide general or ambient lighting with the addition that some of the light can be emitted upward to produce a higher ceiling luminance than recessed or surface-mounted downlights. Examples include fluorescent troffers, compact fluorescent downlights, incandescent and tungsten-halogen downlights for task lighting on kitchen counter tops, and wrap-around lens luminaires (Figure 7-16). Wall Washer. These luminaires are used to produce a distribution of illuminance/luminance on a wall that, though not necessarily uniform, changes gradually from high values at the top of the wall to lower values down the wall. Many wall-wash luminaires are designed to achieve an illuminance ratio from the top to the bottom of the wall of 10:1 or less. Wall-wash luminaires can be recessed or surface mounted. These can be grouped by size. There are two basic types of wall washers (Figure 7-17). Linear fluorescent wall washers usually have a reflector that allows them to be placed close to the wall, and are available recessed or surface-mounted. Other wall washers, including compact fluorescent, incandescent, halogen, or compact metal halide lamp luminaires, are smaller units that, if recessed, have a modest aperture and therefore can appear like other downlights in the space. They also can be surface mounted.

Figure 7-17. Examples of wall-wash luminaires: (a) compact fluorescent lamp luminaire with baffles, (b) incandescent lamp luminaire with eyelid, (c) recessed linear fluorescent lamp, (d) continuous linear fluorescent lamp with baffles, and (e) incandescent lamp wall washer with spread lens.

Figure 7-18. Examples of accent luminaires: (a), (b), and (c) recessed adjustable accent luminaires for reflector incandescent lamps, for tungsten-halogen lamps, and HID lamps; (d) and (e) front and side view of wall sconce luminaires using compact fluorescent and incandescent lamps.

Figure 7-19. Examples of track luminaires: (a) close-up of multicircuit track, (b) and (c) track-mounted luminaires with optical control, and (d) and (e) track luminaires for holding and aiming lamps. Accent. These luminaires are either themselves ornamental or are designed to produce patterns of light that are ornamental. They can be ceiling recessed or surface mounted, or wall mounted (Figure 7-18). Ceiling-mounted accent luminaires use incandescent, tungsten-halogen, compact fluorescent, or low-wattage metal halide lamps. The lamps are adjustable or fixed. Sconces and other wall-mounted accent luminaires use incandescent, tungsten-halogen, or compact fluorescent lamps. Since they are often mounted low, they are often in the field of view, and therefore the designer should be aware of the potential for glare. Translucent shields, which vary in size or shape, are often used for lighting hallways, stairways, doorways, and mirrors. Wall-mounted luminaires with opaque shielding completely conceal the source from normal viewing angles and are strongly directional in light distribution. Downlight luminaires often are mounted on the wall for accent and display lighting, whereas uplight luminaires can be used for general, indirect lighting. The extent to which wall-mounted luminaires protrude from the wall is often subject to code restrictions such as the Americans with Disabilities Act.1 Track. This refers to a system that includes luminaires and a track or rail that is designed to both provide mounting and deliver electric power (Figure 7-19). Track is generally made of linear extruded aluminum, containing copper wires to form a continuous electrical raceway. Some varieties can be joined or cut, and others set into a variety of patterns with connectors. Track is available in line or low-voltage, with remote transformers available for the low-voltage equipment. Track can be mounted at or near the ceiling surface, recessed into the ceiling with special housing or clips, or mounted on stems in high-ceiling areas. It also can be used horizontally or vertically on walls. Mechanical considerations may limit certain mounting arrangements, particularly for wall-mounted installations.

Track can be hardwired at one end or anywhere along its length. Flexibility can be added with a cord-and-plug assembly to supply power rather than with hardwiring. A variety of adjustable track-mounted luminaires are available for attachment at any point along the track. These luminaires come in many shapes and styles, housing a large assortment of lamps, including line and low-voltage. In addition, a number of luminaires are designed to create special effects for decorative applications. Track luminaires use incandescent, tungsten-halogen, compact fluorescent, metal halide, or high-pressure sodium lamps. Point Indirect. These luminaires are designed to provide general or ambient lighting by illuminating the ceiling with compact fluorescent, metal halide, or even high-pressure sodium lamps (Figure 7-20). They contain reflectors that help produce a wide distribution so that they can be mounted close to the ceiling. Pendants or cable usually suspends them, but some types are post-mounted from the floor. Linear Indirect. These luminaires are designed to use linear or compact fluorescent lamps to provide general or ambient lighting by illuminating the ceiling (Figure 7-20). Reflectors are used to produce wide distributions and permit short suspension distance. Linear indirect luminaires can be suspended from the ceiling by pendants or cable or, in the case of modest spans, mounted by their ends. They can also be mounted on the walls to form a perimeter lighting system. Suspended linear indirect luminaires usually have a luminous intensity distribution that is symmetric about the lamps' axis, whereas wall-mounted linear indirect luminaires typically have an asymmetric distribution.

Figure 7-20. Examples of indirect luminaires: (a) pendant-mounted point indirect luminaire with metal halide lamp, (b) and (c) linear two-lamp fluorescent indirect luminaires, (d) linear single fluorescent indirect luminaire, (e) linear two-lamp fluorescent luminaire, and (f) floor-mounted point indirect luminaire.

Figure 7-21. Examples of cove luminaires: (a) cove forming luminaire with biaxial fluorescent lamp, (b) and (c) cove forming luminaires with linear fluorescent lamps, and (d) fluorescent lamp strip luminaire with asymmetric reflector for mounting in a cove. Linear Direct-Indirect. These luminaires are similar to the suspended indirect but provide some downward directed light. Variations are available for changing the proportion of upward and downward light. Cove. These luminaires are designed to be placed in an architectural cove or to have a shape such that when mounted on the wall they produce a cove and its lighting effect (Figure 7-21). The simplest form of this luminaire is a fluorescent lamp strip, providing ballast and lamp sockets. More elaborate forms provide reflectors to control near-wall and ceiling luminance. Stage. These luminaires are designed to produce tight optical control and provide maximum flexibility (Figure 7-22). They are common in theaters and television studios for lighting stage sets and people. Industrial Linear Fluorescent. These luminaires are often designed for high-output fluorescent lamps, with the reflector often being part of the housing (Figure 7-23). A refractor or lens is uncommon. These luminaires are designed to minimize accumulation of dirt by providing for convection; in areas with large amounts of airborne particles, dust-tight covers are used. Diffusers with gasketing are often used in wet locations. Strips. These luminaires have one or more fluorescent lamps mounted to a small housing large enough to hold ballasts and sockets (Figure 7-24). Reflectors are uncommon since these luminaires are used in areas where a large amount of general diffuse lighting is required and efficiency and budget are a concern.

Figure 7-22. Examples of theatre luminaires: (a) Fresnel spot, (b) ellipsoidal spot, and (c) border spot.

Figure 7-23. Examples of linear fluorescent lamp industrial luminaires with (a) parabolic reflector, (b) specular reflector and shield, (c) diffuser and gasketing for wet locations, and (d) open enamel reflector.

Figure 7-24. Examples of fluorescent lamp industrial strip luminaires: (a) tandem lamp strip, (b) two-lamp channel strip, (c) side-mounted lamps, and (d) staggered lamps for continuous row applications. High Bay. These luminaires use HID lamps to produce general lighting in an industrial area (Figure 7-25). They are for applications with spacing-to-mounting height ratios of up 1.0. They are surface or pendant mounted, depending on the structure and openness of the area. These luminaires use reflectors and refractors to produce luminous intensity distributions that vary from narrow to wide, depending on the application and the need for vertical illuminance. Low Bay. These luminaires use HID lamps to produce general lighting in an industrial area (Figure 7-26). They are for applications with spacing-to-mounting height ratios greater than 1.0. As with high bay luminaires, they are surface or pendant mounted. These luminaires usually have wide luminous intensity distributions to provide greater horizontal and vertical illuminances in areas with restricted ceiling heights. Emergency and Exit. Emergency lighting luminaires are designed to provide enough light for egress in emergent situations. They may operate from power provided by batteries. Under normal conditions the batteries are continuously charged from line voltage. These luminaires contain circuitry that turns them on whenever line voltage is not present. Some HID luminaires contain auxiliary light sources that provide light while HID lamps cool off and restrike when line voltage is temporarily disrupted. Exit luminaires help building occupants identify directions to an exit. They can be considered a type of illuminated signage that is useful under normal conditions but designed to provide critical help in emergent situations. Like emergency lighting luminaires, exit luminaires often operate on batteries. Compact fluorescent lamps and light-emitting diodes are commonly used in exit luminaires. Examples of emergency and exit luminaires are presented in Figure 7-27.

Figure 7-25. Examples of high-bay industrial luminaires: (a) open metal reflector, (b) open injection molded acrylic reflector/refractor, and (c) and (d) enclosed reflector.

Figure 7-26. Examples of low-bay industrial luminaires.

Figure 7-27. Examples of emergency and exit luminaires. Outdoor Floodlight. These luminaires are often used for building lighting and other special applications (see Chapter 20, Sports and Recreational Area Lighting, and Chapter 21, Exterior Lighting). These applications can require luminous intensity distributions that range from very narrow to very wide, depending on the angular size of the object being illuminated and the effect to be achieved. The luminous intensity distributions usually are not symmetric. Most types of HID lamps are used in floodlight luminaires. Lamp orientation and reflector arrangement normally determine beam characteristics. Examples of floodlight luminaires are presented in Figure 7-28. Exterior building lighting uses luminaires with narrow and wide distributions, depending on the portion of the building being illuminated and its distance from the luminaire mounting location. Column lighting, accent lighting and distant mounting locations require narrow distributions. Lighting large areas with near mounting locations requires very wide distributions. Floodlight luminaires often have luminous intensity distributions that produce a square or rectangular illuminance pattern (see Chapter 21, Exterior Lighting). Sports Lighting. Some sport lighting luminaires have very narrow luminous intensity distributions and typically are mounted to the side and well above the playing area. Others have medium distributions and sharp cutoff and are mounted over the playing area. Metal halide lamps are common for sports lighting luminaires. Reflectors are used to produce the required luminous intensity distribution. Refractors are not used. Use of the narrow-intensity-distribution luminaires almost always requires careful design to ensure proper overlapping of beams as well as proper horizontal and vertical illuminances. Since aiming is a critical part of their application, these luminaires are usually provided with special aiming and locking gear. Internal or external louvers also may be provided to control glare and light trespass and to improve observer comfort. Examples of sports lighting luminaires are shown in Figure 7-29.

Figure 7-28. Examples of floodlight luminaires.

Figure 7-29. Example of sports lighting luminaires: (a) wide distribution with lamp base to the side, and (b) narrow distribution with hood and baffles to control spill light

Sports lighting luminaires are usually classified using the NEMA field angle designation. Seven categories from very narrow to very wide are used to describe the luminous intensity distribution of these luminaires (see Figure 20-10 in Chapter 20, Sports and Recreational Area Lighting). Street and Roadway. These luminaires are designed to produce reasonably uniform illuminance on streets and roadways. They usually are mounted on arms on a pole, or are post-top mounted. All types of HID lamps are used in street and roadway luminaires. Low-pressure sodium lamps are used only occasionally. Reflectors and refractors are used to produce the various types of luminous intensity distributions required in these applications. Wide distributions permit large pole spacing. Minimum horizontal illuminance and uniformity of horizontal illuminance are typical design criteria. For this reason, the luminous intensity distributions can sometimes have maximum values at angles above 75° from the nadir. Luminaires with dropped-dish, or ovate, refractors are commonly used in roadway applications. Because of their appearance these luminaires are referred to as "cobra head" luminaires (Figure 7-30). Poles for roadway applications are usually mounted well back from the roadside to avoid damage to both the luminaire and oncoming traffic. Modifications to the typical design of cobra heads may reduce glare and light trespass. Pathway. Walkway and grounds lighting is often accomplished with bollards (Figure 7-31). These luminaires are mounted in the ground and have the form of a short thick post similar to that found on a ship or wharf, hence the name. The optical components are usually at the top, producing an illuminated area in the immediate vicinity. Bollards are used for localized lighting. Their size is appropriate for the architectural scale of walkways and other pedestrian areas. Small sharp cutoff luminaires are also used on small poles to provide pathway lighting. Additionally, luminaires for lighting outdoor stairs and ramps are used. These can be mounted on poles or recessed into the structure near the stairs or ramp. Parking Lot and Garage.2 Parking lot lighting often uses cutoff or semi-cutoff luminaires with flat-bottomed lenses. These luminaires are mounted on post-top brackets or on short arms and can be arranged in single, twin, or quad configurations. Symmetric and asymmetric intensity distributions and mounting configurations are used to provide the necessary flexibility in pole placement for parking lots. Examples of parking lot and garage luminaires are presented in Figure 7-32. Wall-mounted luminaires are often used for small parking lots immediately adjacent to a building or in parking structures. Often referred to as "wall packs," wall-mounted luminaires have an asymmetric distribution necessary for lighting adjacent parking lots. There is significant potential for glare and light pollution with these luminaires. Glare and light pollution can be controlled using cutoff versions of wall packs, but this decreases luminaire spacing.

Figure 7-30. Example of a roadway luminaire: a cobra head roadway luminaire with drop-dish refractor.

Figure 7-31. Examples of pathway luminaires: (a) and (b) post-top luminaires for pathway and area lighting, (c) and (d) bollards for pathway lighting, and (e) recessed stair and ramp luminaire.

Figure 7-32. Examples of parking lot and garage luminaires. Surface-mounted luminaires in parking structures are mounted on walls or ceilings. These are designed to produce a considerable amount of interreflected light in the structure. Security. Security luminaires are typically outdoor luminaires designed to help visually secure an area. This can mean providing sufficient illuminance for visual surveillance or security camera surveillance. These luminaires are typically mounted in inaccessible places and have particularly strong housings and lenses to help make them vandal proof. In conjunction with some security camera systems, infrared (IR) sources can also be used that are invisible to potential trespassers. Examples of security luminaires are presented in Figure 7-33. Landscape. Landscape luminaires are designed for use outdoors to light buildings, planting, water features, and walkways (Figure 7-34). The can be mounted in the ground, on poles, on trees, or underwater. Typically they have special housing, gasketing, lenses, and electrical wiring hardware that protects against the effects of water and corrosion.

Figure 7-33. Examples of security luminaires: (a) wall-mounted HID lamp luminaire, (b) wall-mounted metal halide lamp luminaire, (c) compact fluorescent lamp jelly jar luminaire, and (d) HID lamp luminaire.

Figure 7-34. Examples of landscape luminaires: (a) ground and path luminaire, (b) and (c) direct burial and well-mounted landscape luminaires, (d) bollards for lighting pathways, and (e) underwater luminaire.

LUMINAIRE PERFORMANCE Luminaire performance can be considered a combination of photometric, electrical, and mechanical performance. Photometric performance of a luminaire describes the efficiency and effectiveness with which it delivers the light produced by the lamp to the intended target. This performance is determined by the photometric properties of the lamp, the design and quality of the light control components, and to some extent any auxiliary equipment required by the lamp. The electrical performance of a luminaire describes the efficacy with which the luminaire generates light and the electrical behavior of any auxiliary equipment such as ballasts. Luminaire efficacy is determined by lamp efficacy and, if present, the ballast and its interaction with the lamp. Electrical behavior, such as power factor, waveform distortion, and various forms of electromagnetic interference, are properties of the lamp and ballast. The mechanical performance of a luminaire describes its behavior under stress. This can include extremes of temperature, water spray or moisture, mechanical shock, and fire.

Components of Photometric Performance Luminaire Photometric Report. Luminaire photometric performance is summarized in a photometric report (Figures 7-35 and 7-36). Luminous intensity values are determined from laboratory measurements and are reported as the luminaire's luminous intensity distribution. Electrical and thermal measurements are made and often reported. These include input watts, input volts, and ambient air temperature. In addition, some calculated application quantities are usually reported. These include zonal lumens, efficiency, and coefficients of utilization. See "Luminaire Photometry" in Chapter 2, Measurement of Light and Other Radiant Energy, for a description of measurement procedures, and Chapter 9, Lighting Calculations, for a description of the calculation procedures that produce the application data.

Figure 7-35. Data from an indoor luminaire photometric report. See Figure 11-12 in Chapter 11, Office Lighting, for a full indoor luminaire photometric report.

Figure 7-36. Examples of outdoor luminaire photometric reports: (a) and (b) floodlight reports, (c) zonal lumens for a floodlight, and (d) coefficients of utilization for a street lighting luminaire. Components of Luminaire Photometric Reports Luminous Intensity Distribution. The luminous intensity distribution of a luminaire specifies its light distribution characteristics. Luminous intensities in various directions are specified in an angular coordinate system appropriate for the luminaire and its customary application. Most luminaires have luminous intensity distributions specified by luminous intensity values in directions given by angles in a spherical coordinate system. For indoor luminaires, the origin is down (nadir) (Figure 7-37). This is Type C photometry. The elevation (vertical) angle θ has the range 0° ≤ θ ≤ 180°. The azimuthal (horizontal) angle ψ has the range 0° ≤ ψ ≤ 360°. For some outdoor luminaires, usually floodlights, the origin is the primary aiming axis (Figure 7-38). This is Type B photometry. In this case the range of the two angles is -90° to 90°. For indoor luminaires, the range of elevation (vertical) angles, θ, depends on the distribution of the luminaire. The range is usually 0° ≤ θ ≤ 90°, 90° ≤ θ ≤ 180°, or 0° ≤ θ ≤ 180°, depending on whether the luminaire emits light only downward, only upward, or both. Increments of 5° in θ are usually reported, though smaller steps are usually measured and sometimes reported if the luminous intensity distribution changes rapidly with elevation angle.

Figure 7-37. Coordinate system for indoor luminaire photometry; Type C.

Figure 7-38. Coordinate system for outdoor luminaire photometry; Type B. Indoor luminaires that exhibit axial symmetric distributions have luminous intensity reported for ψ = 0°. An incandescent downlight with lamp base up is a luminaire with an axially symmetric distribution. If the luminaire exhibits quadrilateral symmetry in the azimuthal angle, ψ, it is customary to report luminous intensity values for 0° ≤ ψ ≤ 90°. A fluorescent troffer with a prismatic lens is a luminaire with a quadrilaterally symmetric distribution. If the luminaire exhibits bilateral symmetry in ψ, then data are reported for 0° ≤ ψ ≤ 180°. A wall-mounted fluorescent indirect is a luminaire with a bilaterally symmetric distribution. In all cases the increments in ψ are usually 22.5°. For outdoor luminaires the range and increments are variable, the limits of each depending on the angular size of the beam. The luminous intensity values reported for a luminaire are almost always from relative photometry, that is, lamps in the luminaires are assumed to be emitting their rated lumens. Light loss factors can be applied to account for actual field conditions. The measurements are always far-field, that is, the distance at which measurements are made is large enough to consider the luminaire to be a point source. It is assumed that all of the luminaire lumens are emitted from the luminaire photometric center. This point is usually at the center of the opening of the luminaire, in the center of its lens, or at the geometric center of its lamps. For many small luminaires, such as incandescent and fluorescent downlights, far-field measurements are not an issue. This is true also when the distance between luminaire and illuminated point is large compared to luminaire dimensions, as in flood lighting. But for large luminaires located near to illuminated surfaces, calculating illuminances with these luminous intensity values must be done with care. Examples of this situation are under cabinet luminaires or task lights. For more information see "Photometry as the Basis for Calculations" in Chapter 9, Lighting Calculations. In either case, the luminous intensity distribution always gives a general idea of how light is distributed by the luminaire. A convenient way to convey this information graphically is to produce a polar plot of the luminous intensity values. The azimuthal (horizontal) angle in the spherical coordinate system is kept fixed and the elevation (vertical) angle is allowed to move from 0° to 90° or to 180°, with the luminous intensity value at each elevation angle being plotted. This data line represents one plane of luminous intensity distribution data. Similar data lines can be plotted for other planes. Cutoff, uniformity of illuminance, and light patterns can be inferred from such plots. For indoor luminaires, luminous intensity distributions are reported usually in two ways: as an array of values and as a polar plot. In the polar plot, luminous intensities in an azimuthal plane are plotted with a single line, labeled with the azimuthal angle or the plane's orientation. Each azimuthal plane is plotted as a separate line (see the polar plot in Figure 7-35). For outdoor luminaires, luminous intensity distributions are usually reported in Cartesian plots. Luminous intensities in horizontal and vertical planes are reported (see the section labeled "Cartesian Plot" in Figure 7-36a). Average Luminance in Various Viewing Directions. The definition of luminance can be extended to determine the average luminaire luminance, Lave:

where I(θ,ψ) is the luminous intensity from the entire luminaire in direction (θ,ψ) and A′ is the luminous area of the luminaire visible from that direction. This luminance gives a general idea of the luminaire's luminance and appearance but is meaningful only if the luminaire is homogeneous. In this case Lave can be used to assess the potential for discomfort glare. If the luminaire exhibits large inhomogeneities in luminance, this value can significantly underestimate the luminance of some parts of the luminaire. Average luminance is reported in indoor luminaire photometric reports. Zonal Lumens. For indoor luminaires, nested conic solid angle cones can be established with apexes at the luminaire photometric center. Given the size of these cones and the luminous intensity values in them, the number of lumens in each cone can be determined. Each cone defines a conic zone, and the lumens within each are the luminaire zonal lumens. Any azimuthal asymmetry presen