Development of Phosphors for LEDS by L

Development of Phosphors for LEDS By L. S. Rohwer and A. M. Srivastava

ight emitting diodes are based on total U.S. energy use in the building sec- ping cost at startling rates, as shown in semiconductors, which emit spon- tor and 21% of the total electricity con- Fig. 1.3 For red LEDs, the luminous effi- taneous radiation under suitable sumed in the U.S. as a whole. Fluorescent ciency (lm/W) has improved at a rate of forward bias conditions. The basic lighting accounts for 67% of this total. 30X per decade, while the cost has L principle is based on low field injec- Solid-state lighting (SSL)-LEDs have the dropped at a rate of 10X per decade. tion electroluminescence where light is potential for significantly improving Another significant accomplishment is generated when injected minority carri- lighting efficiency leading to a reduction the availability of blue LEDs as a result of ers recombine with the majority carriers in lighting energy use and a concurrent advancements in nitride materials. The radiatively across the band gap of the reduction in pollution from fossil fuel technology continues to progress with crystal. Thus, electrons are injected into power plants. A recent study predicts that the announcement of a 100 lm/W p-type semiconductors and holes are SSL-LEDs with an electrical-to-optical orange/red LED by Lumileds and Philips. injected into n-type semiconductors. The conversion efficiency of 50%, and 200 Lumileds also reported 50 lm/W green physics of LED devices is not dealt with lm/W, could benefit the U.S. by a 50% InGaN LEDs. here and the reader is referred to an arti- decrease in the 600 TW-hr of electricity The technical challenges in SSL can cle by DenBaars.1 per year used for lighting. This translates be grouped into three primary technolo- The recent burgeoning interest in the into cost savings of $25B/year. However, gies:2 field of LEDs is traceable to the break- significant technical challenges must be • Substrates, Buffers, and Epitaxy through in the epitaxial growth of overcome to meet these efficiency goals. • Physics, Processing, and Devices InGaN compound semiconductor mater- These challenges create opportunities for • Lamps, Luminaires, and Systems ial which has led to the development of national laboratories, universities, and blue light emitting diodes (and blue laser industries to work together to develop Luminescent materials (phosphors) sig- diodes). The semiconductor is an effi- the science and technology base required nificantly impact the third technology cient blue light emitter with emission to make SSL-LEDs a reality for general area. The function of a phosphor in SSL- maximum near 450 nm to 470 nm. illumination. Table I shows the targets for LEDs is to absorb the primary wave- There is growing commercial use of blue LEDs set for the year 2002, 2007, 2012, length emitted by the LED chip (near-UV LEDs based on the GaN semiconductor and 2020.2 Achieving the targets for 2007 or blue light) and convert it to visible technology in such diverse applications would enable SSL-LEDs to compete with luminescence. This paper focuses on the as full-color flat-panel displays, traffic incandescent bulbs, and the 2012 targets role of phosphors in solid-state lighting. signals (where they replace the short would enable SSL-LEDs to compete with The requirements of phosphors for lived and energy-hungry incandescent fluorescent lamps. By 2020, if the targets use in SSL-LEDs include: high absorption lamps), and high-resolution printing, to are met, SSL-LEDs are expected to pene- in the near-UV to blue spectral region, name a few. trate all the lighting markets. high quantum efficiency, color rendering According to the U.S. Department of Over the past four decades, LEDs have index (CRI) > 80, resistance to elevated Energy, lighting accounts for 14% of the been increasing in efficiency and drop- temperature (up to 150°C), and mainte-

Table I. Technology performance targets of SSL-LEDS.2

Technology SSL-LED SSL-LED SSL-LED SSL-LED Incandescence Fluorescent 2002 2007 2012 2020

Luminous Efficacy (lm/W)...... 25...... 75 ...... 150 ...... 200 ...... 16 ...... 85 Lifetime (hr)...... 20,000 ...... >20,000 ...... >100,000...... >100,000...... 1,000...... 10,000 Flux (1m/lamp)...... 25...... 200 ...... 1,000...... 1,500...... 1,200...... 3,400 Input Power (W/ lamp) ...... 1...... 2.7 ...... 6.7...... 7.5...... 75 ...... 40 Lumens Cost ($/klm) ...... 200...... 20 ...... <5 ...... <2 ...... 0.4...... 1.5 Lamp Cost ($/lamp) ...... 5...... <5 ...... <5 ...... <3 ...... 0.5...... 5 Color Rendering Index (CRI) ...... 75...... 80 ...... 80 ...... 80 ...... 95 ...... 75

36 The Electrochemical Society Interface • Summer 2003 FIG. 1. Performance in lumens/watt of light sources is shown as a function of time in history. Milestones in materials development and performance are indi- cated. Luminous efficiencies of incandescent and fluorescent lighting tech- nologies are plotted for comparison.

It is desirable to produce a white light solid-state lighting source that has the potential of replacing “common” lamps such as the linear and compact fluorescent lamp that are in wide use for general and specialized illumination applications. To produce white light, the blue emission from the LED can be combined with complementary yellow emission that may be generated by a phosphor. The high absorption of the blue LED light and the efficient conversion into photons of the desired wavelength are the principal determinants of the phosphor’s ability to generate white light in the blue LED + phosphor combination. The yellow luminescence emission with the added transmitted blue light, when λ λ carefully matched, can produce white FIG. 2. Room temperature excitation (left, em= 570 nm) and emission spectra (right, ex = 450 nm) of 3+ Y3Al5O12:Ce . light with color temperatures ranging from 3000 K to greater than 6000 K. The green-yellow emission of Ce3+ in members in the garnet family, nance of high quantum efficiency in an 2 electron volts. This single factor (Y,Gd)3(Al,Ga)5O12, meet most require- encapsulating matrix. Quantum efficien- accounts for nearly 55 percent of the ments specified for this application. cy is defined as the ratio of the number energy loss in the conventional fluores- The optical transitions on the activator of photons emitted to the number of cent lamp. Therefore, the efficiency of a Ce3+ ion are of the allowed 4fn ↔ 4fn-15d photons absorbed. The quantum effi- lighting device would increase as the type. Since the 5d electron interacts ciencies of fluorescent lighting phosphor ultraviolet wavelength is moved closer strongly with its crystalline environ- compositions are in the range: 80-100%. to the average visible light wavelength. ment, broad emission and absorption Fluorescent lamp phosphors convert For example, the energy conversion effi- spectra are obtained. Also, the spectral the ultraviolet discharge of rare-gas/mer- ciency of a 370 nm UV and a 450 nm distribution of emission of the phos- cury (primarily at 254 nm) plasma into blue radiator into visible light would be phor is a strong function of the compo- visible light. The efficiency of conver- about 65 percent and 80 percent, respec- sition and structure of the phosphor sion of primary UV energy into visible tively. Consequently, the development host lattice. Consequently, the emis- light of wavelength 555 nm (this average of a high-efficiency, low cost source of sion color can be conveniently adjusted wavelength of the visible light is fixed by 370-450 nm radiation would be highly by changing the composition of the the human spectral response) in the con- beneficial to general purpose illumina- phosphor lattice. The lowest energy ventional fluorescent lamp is approxi- tion. Two such commercially available Ce3+ 4f5d excitation transition in mately 45 percent. The rather low ener- devices are UV (emission at 370 nm) and Y3Al5O12 (YAG:Ce) is capable of strong- gy conversion efficiency arises from the blue (emission near 450 nm to 470 nm) ly absorbing the emitted blue light fact that each individual ultraviolet pho- LEDs. It should be noted, however, that from the LED (Fig. 2). The emission of ton incident on the phosphor carries the emission intensity of InGaN LEDs YAG:Ce is efficient and consists of a nearly 5 electron volts of energy, while decreases very quickly at emission wave- broad band centered at about 580 nm each visible photon in a mercury-based lengths less than 400nm. (Fig. 2). The complementary blue (from discharge carries (on average) barely over the LED) and yellow (from the phos-

The Electrochemical Society Interface • Summer 2003 37 phor) emission bands offer a resultant white emission. By careful control of the chemical composition of the garnet phosphor type, the Ce3+ emission can be tuned from 510 nm to 590 nm. Consequently, white light with color temperatures ranging from 3000 K to greater than 6000 K can be produced. The hybrid approach (blue LED +YAG:Ce phosphor) is used in most of the white LED flashlights sold commercially. Users of these flashlights will notice a blue halo surrounding the white light beam. This is a result of the combination of two types of light emit- ters, directional (LED) and isotropic (powder phosphor). Because of this, the hybrid approach has limitations. It is also possible to tune the semi- 2+ 2+ 2+ FIG. 3. Room temperature emission spectrum of a bi-phosphor (Sr4Al14O25:Eu and Sr2P2O7:Eu , Mn ) conductor emission to 370 nm by λ white light emitting blend ( ex = 400 nm). changing the indium content in Ga1-x InxN. In such devices, white light can be produced by long-wave UV excitation of a blend of three phosphors, each centered at about 580 nm. The enhance- duce white light by combining the blue of which emits one of the three prima- ment of the Mn2+ emission and the LED emission with a blue excited com- ry colors: red (at 610 nm), green (at 555 quenching of the Eu2+ emission are due plementary green-emitting and red nm) and blue (at 450 nm). Phosphors to the efficient Eu2+Mn2+ energy trans- emitting phosphor.5 Suitable phosphors based on the narrow line emission of fer. The phosphor exhibits strong for such a device are: (1) green emitting, 3+ 2+ trivalent rare earth ions like Eu (red) absorption throughout the UV and SrGa2S4:Eu and (2) red emitting and Tb3+ (green) and broad band emit- almost into the visible. Hence, the phos- SrS:Eu2+. The optical transitions on the ters such as Mn2+ (green) can supply phor luminesces a bright yellow under activator Eu2+ ion are of the allowed red and green components and in com- short and long-wave UV. 4fn ↔ 4fn-15d type. By carefully control- bination with a suitable blue emitting It is noted that supplementing the ling the amount of each phosphor in the phosphor can produce the white light yellow emission of (Sr1-x-yEuxMny)2 blend, white light devices with various with high efficacy. In the “triphos- P2O7 phosphor with a blue-green emit- color temperatures can be achieved. phor,” the amount of each color phos- ting phosphor should produce a white Research aimed at developing new phor can be easily adjusted to produce field. The efficient emission of the com- phosphors that are superior to existing white field corresponding to a wide mercial SAE phosphor in the blue-green phosphors for use in LEDs has received range of color temperatures with high at about 490 nm (quantum efficiency of much attention. Among the new phos- brightness and excellent color render- 90 percent) is particularly attractive phors that have been investigated are ing index. It is, of course, necessary that since the phosphor is well excited by members of the M2Si5N8 (M = Ca, Sr, Ba) the phosphors strongly absorb the 370 short and long-wave UV and has little family of materials activated with diva- nm long-wave UV. Since a majority of or no selective absorption of the visible lent europium.6 Divalent europium-acti- conventional phosphors do not absorb light. The phosphor blend generates vated M2Si5N8 materials display strong well at these longer wavelengths, it is white light with color points in close absorption in the blue (and into the vis- necessary to custom-make new phos- proximity of the black body locus when ible) and exhibit efficient emission in phors to meet the requirements of this energized by line wavelength UV radia- the red. application. tion. It is clear that the motivation of The conversion of short and long- The generation of white light by the many members of the L&DM Division is wave ultraviolet (emission wave- UV LED/phosphor technology has one to develop a triphosphor blend similar length: 370-410 nm) LED radiation by very desirable property (just like in fluo- in concept to the triphosphor blend of a blend of two phosphors to produce a rescent lamp technology and unlike blue fluorescent lamps, which, when ener- white field corresponding to low color LED/phosphor). It affords very precise gized by the LED long UV radiation, will temperature, high CRI and high control over the color of the resultant produce white light suitable for general- brightness, is a subject of a U.S. white light so that it is possible to devel- purpose illumination. One of the patent.4 The phosphor blend is com- op LEDs with different spectral output research activities involves investigating posed of two phosphors, a yellow for general and specialized applications. how to improve the quantum efficiency 2+ 2+ emitting Sr2P2O7:Eu , Mn phos- It is also possible to achieve the of phosphors for SSL-LEDs. These phor and a blue-green emitting phos- “triphosphor” scheme of white light research activities focus on understand- phor such as commercially available generation in the hybrid blue LED and ing the materials properties that influ- 2+ Sr4Al14O25:Eu (SAE). phosphor combination. The strong exci- ence quantum efficiency and develop The efficient blue emission at 420 tation/absorption spectra of some diva- methods to enhance QE through novel nm of divalent europium (Eu2+) in lent and trivalent rare earth activated sensitizers, precise control of particle Sr2P2O7 is well known. The additional sulfides extend well into the visible and size, and surface chemistry. An addition- incorporation of Mn2+ produces a very exhibit resonance with the blue LED al concern of phosphor researchers is the efficient yellow emission band that is radiation. It has been proposed to pro- maintenance of high quantum efficien-

38 The Electrochemical Society Interface • Summer 2003 cy at elevated temperatures in an encap- Acknowledgments 5. G. O. Mueller and R. Mueller-Mach in sulating matrix. The high input power Physics and Chemistry of Luminescent Materials, C. R. Ronda, L. E. Shea, and A. M. densities and temperatures generated by The contributions of Jeff Tsao and Jerry the LED chip affect the phosphors, Srivastava, Editors, PV 99-40, p. 91, The Simmons of Sandia National Electrochemical Society Proceedings Series, which are in close proximity to the chip. Laboratories are appreciated. Sandia is a Pennington, NJ (2000). Heat is transferred to the phosphor multiprogram laboratory operated by 6. R. B. Mueller-Mach, G. O. Mueller, T. Juestel through an encapsulating matrix such as Sandia Corporation, a Lockheed Martin and P. Schmidt, U.S. Pat. 20030006702 a silicone gel. Phosphors could experi- Company, for the United States (2003). ence temperatures up to 150°C in real Department of Energy under contract devices. DE-AC04-94AL85000. The development of highly efficient and inexpensive LED-based white light References About the Authors lamps would certainly represent the most significant revolution in lighting 1. S. DenBarrs, in Solid State Luminescence; LAUREN SHEA ROHWER is the Vice- technology since the invention of the Theory, Materials and Devices, edited by A. H. Chair of the L&DM Division and a fluorescent lamp. Benefits clearly exist Kitai, Chapman and Hall, London (1993). Senior Member of the Technical Staff for the lighting industry and beyond. 2. J. Y. Tsao, Ed., in Light Emitting Diodes (LEDs) at Sandia National Laboratories. Achievement of the project goals will for General Illumination Update 2002, She may be reached by e-mail at save money for the end-users, while the Optoelectronics Industry Development Association, Washington, DC (2002); and [email protected]. whole country will benefit from M. Stolka, Ed., in Organic Light Emitting reduced energy consumption, reduced Diodes for General Illumination Update 2002, ALOK SRIVASTAVA is the past Chair of pollution from fossil fuel power plants, Optoelectronics Industry Development the L&DM Division and a Senior and elimination of costly disposal that Association, Washington, DC (2002). Solid State Chemist at GE Global will result in a safer environment. The 3. Figure compiled by J. Y. Tsao from data in Research Center. He may be reached targeted market segments are commer- Ref. 2. by e-mail at srivastava@ crd.ge.com. cial, industrial and retail, all of which 4. A. M. Srivastava and H. A. Comanzo, U.S. are very interested in the higher effi- Pat. 6501100 (2002). ciency products.

The Electrochemical Society Interface • Summer 2003 39