High-brightness LEDs have developed rapidly during the past five years. They are used in traffic lights, signs and specialty lighting and in new applications such as backlight units for LCDs, headlamps for cars and street lighting. The efficiencies of white light from high-brightness LEDs now have reached greater than 100 lm/W, while costs keep decreasing. The efficiency per cost of high-brightness LEDs has been doubling every two years, which will lead to even wider use in general lighting in the future. A high-brightness LED is a solid-state light-emitting chip that emits in blue, green and red wavelengths. Typically, it consists of multiple quantum layers grown between p- and n-doped gallium nitride (GaN), which is grown on a lattice-matched substrate such as silicon carbide or sapphire. The substrate can be removed with a laser liftoff technique, and the GaN can be bonded to silicon. Direct-wire bonding or flip-chip bonding techniques are used to make electrical contacts to the chip. White light For white-light generation, a phosphor layer is added to the blue-emitting GaN chip surface to convert the blue wavelengths (400 to 480 nm) to broadband white light. Finally, a hemispherical lens can be attached optionally on top to give a lambertian pattern of light emission. Today, high-brightness LEDs emit 400- to 480-nm fluxes in excess of several hundreds of milliwatts in a chip with an area of 1 mm2, which is then converted to white light in a phosphor layer. Typically, these devices run on 0.3 to 1 A or more with an input electrical power of 1 to 5 W. Although the LEDs do not radiate heat in the way that an incandescent lamp does, they do generate heat that must be conducted away by a heat sink. The junction temperature of these LEDs can be anywhere from 60 to 180 °C, depending on the chip size, package, current flowing through it, ambient temperature and humidity. Potentially high light fluxes combined with high temperatures put a lot of stress on the materials used to make high-brightness LEDs and are the subject of continuing improvements in chip design, thermal management, packaging and improved polymers for encapsulation. Encapsulants play a key role in the optical performance and packaging of high-brightness LEDs. The 2.5 refractive index of GaN causes the light that is emitted within the multiple quantum wells to be highly confined and wave-guided within the chip if the surface is bounded by air. The escape cone angle of light is bounded by the critical angle (23.5°), which is given by sin θc = n(encapsulant)/n(GaN). Therefore, it is desirable to cover the surface above the GaN with an encapsulant made of a material with a very high refractive index to extract as much light as possible. When light output is calculated as a function of angle for encapsulants of various refractive indices, the higher the refractive index, the larger the escape cone angle and the higher the optical transmission (Figure 1). This calculation is based on a simple plane interface between GaN and the encapsulant. Therefore, it is important to design liquid high-refractive-index encapsulants that can be dispensed in a manufacturing setting and can be cured within a reasonable time and temperature. Figure 1. The optical transmission through GaN is shown with the encapsulant interface as a function of internal incidence angle of rays emitted from the active multiple quantum layer. At and above the critical angle, no light is transmitted from GaN to the encapsulant. RI = refractive index. However, some high-brightness LED researchers and manufacturers also use a combination of surface roughening and encapsulants. Alternatively, photonic bandgap structures can be used to diffract light from the surface. The encapsulant also acts as a binder for the phosphor. Thus, it is desirable for the refractive index of the encapsulant to approach that of the phosphors (typically ~1.7 to 1.8) to minimize backscatter and to enhance light output. Some companies mix the phosphor in the encapsulant, and issues such as polymer viscosity are important to avoid settling of the phosphor during cure. The encapsulant also acts as a filler or adhesive between the phosphor layer and the top hemispherical lens. Designing an encapsulant for high-brightness LEDs is challenging because many — and sometimes conflicting — requirements must be met simultaneously. High optical transparency (>98 percent for 1-mm thickness) is needed at the operational wavelength to minimize residual absorption of light and heating of the encapsulant, which can result in discoloration. Rohm and Haas Electronic Materials of Spring House, Pa., recently developed the XP077171 (refractive index = 1.6 at λ = 633 nm) and XP07286 (refractive index = 1.57 at λ = 633 nm) liquid high-temperature-use encapsulants for high-brightness LEDs (Figure 2). If the refractive index of a polymer is raised, the absorption edge of the transmission curve moves toward the visible — i.e., there is loss of transparency; however, these encapsulants’ optical transparency window is down to λ = 350 nm. Figure 2. The optical transmission of Rohm and Haas encapsulants shows an absorption edge near 350 nm, giving transparency when the wavelength is >400 nm. The thickness of the film is 1 mm. Another key requirement of an encapsulant for high-brightness LEDs is high heat stability because it can experience continuous temperatures as high as 150 to 200 °C in air. Problems from heat manifest as yellowing of the polymer, which can have deleterious consequences for the color chromaticity of the LEDs. The increased absorption of the light output results in premature device failure. The CIE b index value is a measure of the yellowing of polymers. Plotting the measured b values of the new encapsulants versus time at 200 °C showed a stability of longer than three weeks (Figure 3). We also have carried out accelerated aging testing at 190, 230 and 250 °C using a 488-nm laser emission with 400 mW/mm2 power and a sample thickness of 1 mm. Fitting the data to an Arrhenius equation predicted that the polymers will last longer than three years at 130 °C at those high light fluxes. Figure 3. CIE b values (yellowness index) of XP077171 and XP07286 are shown over a period of 25 days when heated to 200 °C. Silicones are particularly suited as encapsulants because of their high thermal stability. For high-brightness LEDs, they are typically two-component systems that are mixed together at the point of use. When the refractive index of polymers approaches 1.6, they become very viscous and difficult to mix and also to dispense. Our encapsulants are one-pot systems and are premixed. They have a viscosity adjusted to a value that makes them easy to dispense. That means that there is no waste or extra bubble formation as a result of vigorous mixing, and the viscosity can be tailored to the requirements of the customer. Currently, two-component systems have limited shelf life (typically eight hours) after mixing; then they must be thrown away, leading to considerable waste. The new encapsulants have a proprietary catalyst that allows them to be stored at 5 to 20 °C for more than six months. The viscosity does not change by more than a factor of two over that period. Our direct measurements of the viscosity over more than three months, as well as accelerated aging tests at elevated temperatures, have shown that the viscosity is stable (Figure 4). Figure 4. The stability of the viscosity of both XP077171 and XP07286 is shown at recommended storage conditions. RI = refractive index. Meet the authors Garo Khanarian is program manager of optical packaging, David Conner and David Mosley are senior scientists, David Thorsen is a scientist, and Ethan Simon is director of microfabrication, all at Rohm and Haas Electronic Materials in Spring House, Pa.; e-mail: gkhanarian@rohmhaas.com.