After 125 years of faithful service, Edison's light bulb is still a primary source of illumination throughout the world. But many in the semiconductor and lighting industries — and even Uncle Sam — would like to retire the incandescent lamp. With an efficacy of 15 lumens per watt (lm/W), incandescents are now the least efficient of the common lamp types. On a grand scale, such inefficiency translates into tremendous energy consumption.
According to one estimate, the amount of energy consumed domestically by incandescents equals the output of 40 large (1000-MW) power plants. Meanwhile, the total energy consumed in this country by all lighting is roughly equal to 100 large power plants, or about one-fifth of our electrical energy consumption.
The good news is that existing lighting technologies have the potential to drastically reduce the energy consumed by lighting. Compact fluorescents, useful as replacements for incandescents, achieve an efficacy of about 60 lm/W. Meanwhile, fluorescent lamps in commercial applications operate around 85 lm/W. These numbers ignore factors such as light quality, lamp reliability and longevity, so the fluorescent-versus-incandescent comparisons are not quite apples-to-apples. But they illustrate the opportunity for energy savings using existing technology, provided the cost of fluorescents can be lowered.
Though fluorescents clearly will play a role in cutting energy consumption, many view solid-state lighting (SSL) — using high-brightness white LEDs or OLEDs — as the ultimate path to more efficient lighting. White LEDs are now achieving as much as 40 to 50 lm/W efficacy. One vendor has even demonstrated 74 W/lm performance in the lab — a level on par with fluorescents.
Going forward, further improvements should boost the output of white LEDs to 150 lm/W. At that level, the LEDs could replace incandescents and fluorescents in general lighting. According to a Department of Energy (DOE) road map, white LEDs should reach this level of performance in the lab sometime around 2012.
It will take additional development effort to bring 150-lm/W white LEDs to market and time for these LEDs to penetrate general lighting applications. A major supporter of SSL, the DOE projects that an actual changeover to SSL in the marketplace could take place by 2025.
Besides the improved energy efficiency afforded by SSL, LEDs offer advantages like greater durability and longer operating life. And LEDs, unlike incandescents, can be dimmed without changing the color of the light they produce. These advantages have spurred the adoption of colored, high-brightness LEDs in signage and in architectural lighting. Meanwhile, white LEDs are proliferating in mobile phones, where they provide backlighting for color LCD displays.
High-brightness LEDs currently account for billions of dollars in sales annually. Nevertheless, even greater potential for SSL lies ahead in general lighting, where the global market for lamps, ballasts, fixtures and lighting controls is measured in tens of billions of dollars. That financial lure has been enough to unite competitors such as GE, Philips Electronics and OSRAM in the pursuit of SSL development and commercialization through their membership in the Next Generation Lighting Industry Alliance.
Naturally, many technical issues must be resolved to make SSL viable. The two main obstacles now are a lack of brightness from existing LEDs and their high cost. Though the challenges are not trivial, researchers believe there are no fundamental barriers to LEDs achieving the performance or cost goals required for SSL.
As SSL continues to develop, there should be many opportunities for the power semiconductor industry to contribute to this revolution in lighting. Given the high-volume nature of the applications, there will be a great need for power ICs that are highly optimized for performance and cost. Chip makers that can exploit process and design expertise to achieve the most integrated power solutions for lighting stand to capitalize on the coming SSL opportunities. Although SSL may still be several years away from a boom, now is the time for power chip vendors to ponder whether their technology road maps and the SSL technology road maps could happily converge.