Since the invention of visible light-emitting diodes (LEDs) based on III-V semiconductor p-n junction materials by Nick Holonyak, Jr., in 1962, LEDs have been developed extensively and are now challenging Edison-style incandescent lamps and are being used in more and more applications in lighting. Their operating wavelengths have decreased over time: first, red LEDs were developed, followed by amber and green LEDs. LEDs entered the market in indicator and signal applications, replacing small incandescent bulbs, thanks to their high efficiency, and hence lower power consumption, and their exceptional reliability. However, extending their operation to even shorter wavelengths was limited by the widest direct-bandgap energy of III-AsP materials, roughly corresponding to a yellow-green color. It took another 30 years for LEDs to cover the whole visible spectrum, when, in 1991, Shuji Nakamura demonstrated commercially viable blue LEDs, based on new wide-bandgap III-N materials. This milestone positioned LEDs as a serious competitor in general lighting, rather than mainly a technology for small indicator lights. The ability of LEDs to emit photons in all three primary colors and to pump phosphors to produce white light marked the beginning of the era of solid-state lighting (SSL).
The levels of efficiency and reliability achieved by LEDs today are by far superior to those of traditional light sources used for lighting. For instance, the peak efficiency of white LEDs exceeds 260 lumens per watt (lm/W), compared with ~16 lm/W for incandescent lamps and <100 lm/W for fluorescent lamps. This level of performance is still growing. SSL based on LED technology will be an important technology for generating significant energy savings and consequently for environmental benefits. However, in order for LEDs to become a main player in lighting, several technical and nontechnical challenges still need to be overcome. The major challenges include further improvements in luminous efficiency while simultaneously delivering superb color quality at reasonable cost, especially for the development of high-power and large-area LEDs with improved light output power and luminous efficiency under high-current injection conditions. While the peak efficiency of the LEDs is exceptionally high, it does not remain high under high-current conditions. This is a well-recognized but not completely understood phenomenon commonly referred to as an “efficiency droop,” which is observed as a reduction in emission efficiency with increasing injection current density. This illustrates the typical concurrent nature of work on visible LEDs, which combines fundamental research, product development, and commercialization of the products—researchers are still exploring the most critical and fundamental device physics issues of the LEDs, while consumers can acquire LED light bulbs at local retail stores. Obviously, ownership cost is the most critical determining factor for successful market penetration of SSL. This cost, however, is not purely an independent economical variable mainly governed by capital cost of the SSL paid by consumers for the adoption of new lighting technology, but rather total cost, heavily dependent on technology advances, consisting of not only capital and operating costs paid by each individual but also environmental cost paid by the whole society. Further improvements in luminous efficiency and reliability as well as reduction in manufacturing cost are required to lower the operating, capital, and social costs of lighting.
The total external quantum efficiency, η ex, of an LED depends on the internal quantum efficiency, η int (a product of current injection efficiency, η inj, and radiative recombination efficiency, η rad), and the extraction efficiency, η ext: so η ex = η int × η ext = η inj × η rad × η ext. In order to improve extraction efficiency of LED devices, photons generated in the active region should escape out of the naturally formed slab waveguide structure formed by the LEDs’ epitaxial layers. In this Focus Issue, several novel approaches for enhancing extraction efficiency are presented. Seong-Ju Park and his colleagues from the Gwangju Institute of Science and Technology and Samsung LED Co. in Korea demonstrate that tungsten metal can be used not only as a mask for epitaxial lateral overgrowth, but also for the formation of an air void underneath it, to improve both the internal quantum efficiency and the extraction efficiency. Huai-Bing Wang and his colleagues from the Suzhou Institute of Nano-Tech and Nano-Bionics of the Chinese Academy of Science take a similar approach: they form pyramidal patterns on sapphire substrates for the improvement of both materials quality of the LED epitaxial structures and of the extraction efficiency. While similar approaches using patterned sapphire substrates have been demonstrated, in this study the resistance to electrostatic discharge effects, which impacts LED reliability, was improved simultaneously.
While internal quantum efficiency has been dramatically improved for blue- and red-emitting LEDs, the efficiency of LEDs emitting in the green at λ = 540–550 nm is still substantially lower: e.g., according to the latest reports: η int > 60% for blue InAlGaN LEDs emitting at λ~460 nm; η int > 90% for red InAlGaP LEDs (λ~650 nm); while η int < 20% for green InAlGaN LEDs (λ~550 nm). This performance deficiency is often referred to as a “green gap” and is associated with a number of fundamental scientific challenges that need to be resolved. For the improvement of the internal quantum efficiency, improvements in material quality and epitaxial layer designs are required to further reduce the density of nonradiative recombination centers and to mitigate the quantum-confined Stark effect (QCSE) in the quantum-well active region, respectively, both of which adversely affect the radiative recombination rate. The epitaxial structures based on III-N materials grown in polar directions on (0001) substrates possess a built-in electrostatic field near the interfaces due to spontaneous and piezoelectric polarization effects. This field is responsible for a reduced overlap of the carriers’ wavefunctions. A paper by Nelson Tansu and his team from Lehigh University describe several strategies and propose optimized epitaxial structures for the mitigation of the QCSE. They tailor the electronic band structure of the layers that form the active region of green-emitting LEDs to improve the electron-hole wavefunction overlap, thereby improving the internal quantum efficiency and the carrier injection efficiency. While Tansu et al. focus on the mitigation of the QCSE in polar III-N structures, Christian Wetzel and Theeradetch Detchprohm from the Rensselaer Polytechnic Institute report on material and device characteristics of QCSE-free nonpolar (10-10) and (11-20) structures, and they demonstrate wavelength-stable green LEDs in order to address technical challenges associated with the polar structures. C. C. Yang and Yean-Woei Kiang’s team from the National Taiwan University investigate surface plasmon coupling with radiating dipoles (electron-hole pairs) experimentally and theoretically. The team demonstrates improvements in the efficiency droop as well as in the internal quantum efficiency. They also numerically study the effects of coupling between a radiating dipole and the localized surface plasmons induced by Ag nanoparticles.
Typical III-N-based visible LED structures are grown on sapphire substrates. While the capability to grow device-quality materials on such substrates with a large lattice and thermal mismatch was a key breakthrough in the development of visible LEDs, sapphire substrates are not an ideal choice. Their low thermal and electrical conductivity are major limiting factors for high-power operation of visible LEDs, which is critically important in SSL applications. Ideally, the sapphire substrate is separated from the epitaxial layer structure after growing the LED material. Hyunsoo Kim and his colleagues from Chonbuk National University, Korea Polytechnic University, and Korea Electronics Technology Institute describe a fabrication process and a LED design with a vertical geometry and without a sapphire substrate. The paper reports in detail on the etching, ohmic metallization, laser lift-off, light extraction, current spreading, and the stability and reliability of the devices. Sapphire substrates are also smaller and more expensive than silicon substrates. Hence, a possible way to further lower the capital cost of SSL technologies based on LEDs is to fabricate the devices on silicon substrates. Kei May Lau and her team from the Hong Kong University of Science and Technology report on blue-emitting LEDs on silicon substrates. The paper addresses many important technical issues associated with LEDs on silicon substrates, such as strain management and crack-formation in the epitaxial structure, thermal management of the chips, and the external quantum efficiency of the devices, including the light extraction.
In order to generate white light from LEDs, several approaches have been explored. The most common approach at present is using blue-emitting LEDs to optically pump (and also to color-mix with) longer-wavelength broadband emitting phosphors. Hao-Chung Kuo and Chien-Chung Lin and their colleagues from the National Chiao Tung University, Tsing Hung University, and the Research Center for Applied Sciences in Taiwan propose a rather simple but effective technique for the improvement of color and temperature stability of such “phosphor-converted” LEDs. The technique is based on the patterning of a remote phosphor packaging using a pulsed spray deposition. Even with their current dominance in phosphor-converted LEDs, YAG:Ce3+ phosphors limit the color rending index of white LEDs to a value of ~80. The color quality has to be further improved to compete with traditional lighting technology. Christopher Summers and Hisham Menkara and their colleagues from PhosphorTech Corp. in USA discuss nano-phosphors based on Mn-doped ZnSeS to enhance the color properties, luminosity, and efficiency of phosphor-converted white LEDs. Another, potentially better, approach for producing white light is based on the mixing of color elements red-green-blue (RGB) or red-yellow-green-blue (RYGB). Improvements in color quality and luminous efficiency are obtained at the expense of the increased cost associated with the mixing of different color elements from individually optimized active devices. Hence, this approach may fit well in high-end lighting systems. Wen-Shing Sun and his team from the National Central University in Taiwan propose an interesting scheme of hybrid light mixing by combining visible daylight and RGB LEDs. This paper also describes a method of color mixing by RGB LEDs with and without sunlight. It is believed that current and future SSL is based on individual primary color LEDs and/or LEDs combined with phosphors. Jeff Tsao and Jonathan Wierer and their colleagues from Sandia National Laboratories, University of New Mexico, and the National Institute of Standards and Technology challenge this common belief that the narrow spectral linewidth and the high capital cost of lasers makes them unsuited for general illumination purposes. They discuss the use of lasers for higher power and efficiency at high current densities for SSL and experimentally demonstrate that four-color (RYGB) laser white illuminant is virtually indistinguishable from high-quality state-of-the-art white reference illuminants. This result suggests that lasers can also be a serious contender for solid-state lighting in some applications.
Ultraviolet (UV) LEDs are also an important light source not only for pumping visible phosphors to produce white light but also for replacing mercury lamps with new applications of air, water, and surface sterilization and bioagent detection. For UV LEDs, AlGaN materials, as opposed to InGaN materials for visible LEDs, are used in active region and the efficiency of UV LEDs are significantly lower than that of InGaN-based LEDs. In order to address one of technical challenges associated with UV LEDs, i.e., low internal quantum efficiency, Kouji Hazu and Shigefusa Chichibu from Tohoku University in Japan experimentally and theoretically investigate the optical characteristics of QCSE-free but anisotropically strained AlGaN materials grown on non-polar (10-10) GaN substrates.
Emerging nanotechnology can also impact the performance of LEDs. JianJang Huang and his team from the National Taiwan University report on nanorod array structures in LEDs and demonstrate that strain and the optical transition energy can be controlled in such structures. Nanorod-LED arrays may play an important role in next-generation visible LEDs as the control of strain can impact the QCSE, which in turn affects the efficiency droop and spectral stability.
The papers in this Focus Issue provide an overview of the current trends and the state-of-the-art in research and development activities in the field of LEDs for SSL. We hope that the readers of Optics Express and its Energy Express supplement will enjoy learning about the latest advances described in this Focus Issue and that it will motivate them to publish their own latest discoveries in the area of optics for energy in Energy Express. We also want to express our sincere gratitude to the contributors who accepted our invitation to contribute an article and who worked to stringent publication deadlines. This Focus Issue, Optics in LEDs for Lighting, would not have been possible without the efforts of Martijn de Sterke (University of Sydney), Editor-in-Chief of Optics Express; Bernard Kippelen, Editor of Energy Express; and the work of the Associate Editors, reviewers, and the staff coordinating OSA’s publications. We want to express our gratitude to all of them.
Atlanta, June 30, 2011