Molecule by Molecule by Molecule
Louis Guido (left) and G.Q. Lu (right) discuss sapphire crystal substrates in the new laboratory for metal-organic chemical vapor deposition (MOCVD). The new lab gives Virginia Tech researchers the capability of building semiconductor materials with precise control of thicknesses and composition at the atomic level.
Louis Guido and G.Q. Lu are trying to build a better light bulb by combining an unconventional approach to the architecture and packaging of light emitting diodes (LED) with nano-technology developed at Virginia Tech. Their effort is now possible with the new ECE/MSE metal-organic chemical vapor deposition (MOCVD) laboratory, which begins operation in June.
Guido and Lu hope to create a white LED lamp with thermal capabilities that are three to five times better than current state-of-the-art. Their efforts support U.S. Department of Energy (DOE) goals of developing solid-state lighting technology that competes with conventional incandescent and fluorescent lamps and is fully integrated into the lighting market by 2020.
DOE statistics for 2001 indicate that lighting accounts for 21 percent of U.S. electric energy generation. Since today’s illumination technology for residential applications still relies on the century-old incandescent lamp developed by Thomas Edison, more than 85 percent of the electrical energy spent on residential lighting is wasted in heating the environment, according to Guido. “This is because incandescent lighting is based on the principle of black-body radiation, which uses electricity to heat a tungsten filament to glow, thus providing light, but also heat,” he explained. “Although fluorescent lighting is about three-to-five times more efficient than incandescent, it has not been widely used in residential markets because of the perceived poor quality of light and higher initial costs.”
By switching to white LEDs, which are efficient and long-lived, national electricity demand for lighting can be reduced by about 167 billion kW hours per year, or the equivalent of about 29 600-MW power plants. The cost savings of $12 billion/year would be augmented by the accompanying reduction in pollution associated with electricity generation.
The p-n junction LED is a nearly perfect electrical-to-optical energy conversion device; that is, it can convert input electrical energy, in the form of the bias voltage applied across the p-n junction multiplied by the current flowing through the device, into output optical energy, consisting of photons created by electron-hole recombination across the energy band-gap of the semiconductor, with an efficiency approaching 100 percent.
The first practical red LED was invented more than 40 years ago, by Nick Holonyak, Jr., who was Guido’s Ph.D. thesis advisor at the University of Illinois. Single-color LEDs are highly efficient and are now used extensively in automobile brake lights, outdoor TV screens, and traffic signals. The challenge for future LEDs in general lighting applications is to generate a “high quality” white light spectrum with the best possible electrical-to-optical energy conversion efficiency. “In principle, the white light output from an LED lamp could be tailored to mimic the solar spectrum over the visible spectrum while eliminating the unnecessary ultraviolet and infrared radiation,” Guido said.
Single-color LEDs used in signaling and display applications now constitute a $2 billion/year industry and the market is expected to grow to $10 billion/year over the next 10 years, he indicated. If white LED lamps could fully penetrate the general illumination market then overall LED sales would double in size beyond this estimate. However, for white LED lamps to successfully compete with incandescent and fluorescent lighting, significant cost improvements are needed. “To meet the roadmap for LED lighting, we need an eight-fold increase in performance, plus a 12-fold decrease in cost,” Guido noted.
“There are two ways to lower the fixed cost of LED based solid-state lighting,” Guido explained. “You can increase the power, or you can lower the cost. It’s hard to increase the power without increasing the cost. A big fraction of the cost is the semiconductor itself, measured in square cm. If you double the power by doubling the area, you have doubled the cost.”
Guido and Lu’s solution is to design the semiconductor nano-scale active region for improved efficiency, then to run more current through, increasing the brightness. Increasing the current will raise the operating temperature, and they are developing the packaging to withstand high temperatures.
“This is a nice collaborative project at the interface between electrical and materials engineering,” said Guido, who, like Lu, has a joint appointment in ECE and Materials Science and Engineering (MSE). “This project involves materials, device design, and packaging.”
Their new design involves removing the sapphire substrate that supports the junction in conventional LEDs so that current can flow directly across the junction. They also plan to use mesh electrodes for optimal light extraction and flexibility, a reflective silver coating, and direct-bond-copper heat sinks.
Increasing the current will yield brighter light per unit area of device, but it will also raise the junction temperature. Conventional packaging materials, such as solder, cannot withstand higher temperatures. The team’s solution is a nano-scale silver paste developed by Lu’s research group working with the Center for Power Electronics Systems (CPES) for attaching power devices. The paste, which looks black at room temperature, turns silver at high temperatures and has excellent thermal and electrical conductivity.
“Our new material works at high temperatures, such as 500-600°C,” Lu said. “However, it can be processed or sintered at the lower 200 to 300°C temperature of solder. Using our paste, manufacturers do not need to retool, which is important for being useful in the marketplace.” He explained that nano-scale materials have a “large surface area,” so the nano-scale paste contains more surface energy than commercial materials and does not need high temperature to attach devices. The sintered silver is also porous, which makes it more malleable, giving devices improved reliability compared to devices attached by conventional means. “With its porosity, the attachment does not transfer stresses from substrate to device,” Lu said.New MOCVD Laboratory enables atomic-level precision
A series of other nano-materials developed at Tech by Lu's group. The yellow and green solutions are suspensions of silver nano-particles, and the rest are suspensions of gold nano-particles. The nano-metal suspensions are under development for biological sensing applications and for control of harmful bacteria.
The third partner on the team to develop a brighter LED is the new ECE/MSE metal-organic chemical vapor deposition (MOCVD) laboratory. The $2.5-million facility, which begins operation June, will enable researchers to build semiconductor materials with precise control of thickness and composition at the atomic level. “We can build very complex structures out of individual layers,” said Guido, who serves as director of the laboratory.
The Aixtron MOCVD system, which is the centerpiece of this new laboratory, can be used to synthesize semiconductor alloys and heterostructures containing elements from column III and column V of the periodic table. These semiconductor materials can be used to construct a variety of devices including electronic amplifiers operating at high power and high frequency and lasers and photodetectors covering the entire ultraviolet-to-infrared portion of the electromagnetic spectrum.
The MOCVD laboratory is the newest addition in the college’s microelectronics capabilities. “The ECE and MSE departments have made a large commitment to microelectronics at Tech,” Guido said. “The capabilities of this laboratory, combined with the soon-to-be-completed upgrade of our cleanroom microfabrication facility, will put Virginia Tech researchers in a position to compete and do research at a new level.” These experimental facilities will enable research along the entire “food chain” from molecules-to-devices in important areas such as chemical and biological sensing, electronics and photonics, and energy and environmental systems.