Despite their efficiency and improved performance, there is still a significant drawback to CFLs -- the light they produce is not as nice as the light from incandescents. Lighting designers use several methods to measure the "quality" of light sources, one of the most important of which is a light's color rendering index, or CRI. The CRI describes the light source's effect on colored objects; the higher a light's CRI, the truer the color of an object under that light.
By definition, a 100-watt incandescent light bulb has a CRI of 100, the highest possible measure, and one that's very close to the color rendering of natural daylight. Fluorescent lamps have low CRIs -- often in the 70-to-85 range. Another important measure of light is its "color temperature," a technical description of how "warm" or "cool" a light appears. Incandescent bulbs have a relatively low color temperature, the kind of comforting, orangish glow people like to have around the house. Newer CFLs can approximate this warmth, but fluorescents have traditionally produced harsher light, a "cold" illumination that offers little comfort.
So while compact fluorescent lights might be a valid alternative to incandescents in some applications -- who needs good color rendering in the bathroom? -- it's unlikely they'll become ubiquitous. "If you think about what makes good light, it depends on where you're using it," says Keith Scott, the business development manager of Lumileds, a Silicon Valley firm that manufacturers high-brightness LEDs.
"In your retail experiences," Scott says, "some of those big-box stores, you'll see fluorescents in the ceiling just spraying light all over the place -- and the lighting that comes down from there is adequate. But when you get into the shopping malls you see people using halogen lights to light clothing, because the halogen light is full spectrum -- you see all the colors. You rarely see people in fine restaurants eating under anything but incandescent or halogen light -- you want to see what your food looks like."
Lumileds is one of the dozens of companies looking to eventually produce what might be the holy grail in lighting -- a practical, efficient, high-quality white LED, something that might become a true replacement for Edison's bulb. There are many steps, and many hurdles, on the way to that goal, though, and for the moment the company is focusing on more conventional markets, one of which is traffic signaling.
A traditional traffic light is unabashedly inefficient, Scott says. "If you think about it, you're putting this 135-watt incandescent bulb behind this red or green or amber filter, and you end up only getting about 10 percent of the light out of there," he says. "You have to crank this monster up just get this small amount of light."
LEDs are monochromatic -- they can produce one color only, meaning that the light from them does not need to be colored by a filter. So a 135-watt bulb can be replaced with 6 or 12 watts of LEDs -- a huge savings. The LEDs also last longer than standard bulbs, so they save municipalities in long-term maintenance costs. For this reason, in large cities, Scott says, the transition to LED signaling is nearly complete -- drive through San Francisco and New York and many of the lights you'll see are powered by LEDs.
In general, Scott says, demand for high-brightness LEDs is currently concentrated in colored-lighting applications -- "lighting at clubs, discos, rock 'n' roll stages, to a lesser extent TV studios. Where we've seen the most success is in mobile stage lighting, like at theater shows and rock 'n' roll concerts."
At last year's Glastonbury Music Festival in England, LEDs were used in one of the 14 concert stages, Scott says. They were so successful there that, at this year's show, which takes place at the end of June, 11 of the stages, including the main stage, will use LEDs. Engineers aren't choosing these lights just because they're cheaper and more durable, but also because they allow for more creative flexibility.
"Because they're digital devices, you can do all the control digitally," Scott says. "What they're doing is they'll look at each light -- imagine this fixture containing red, green and blue LEDs -- and address each color to produce new and different colors. They'll create waves and washes of color onstage, things that are impossible with anything else."
So LEDs are good at color -- but what about white, natural-light LEDs? Because it's not until the LED takes the household light by storm that any appreciable impact on widespread residential electricity consumption will occur.
But that's a little trickier. So far, the only practical white LEDs that engineers have managed to produce is a kind of hack. To make an LED glow white, they coat a blue LED with a "yellow" phosphor -- a compound that absorbs part of the blue light and re-emits it as yellow light. The blue and the yellow light combine to form white, but it's not a very good white light, Scott says -- the light often has a very harsh temperature and poor color rendering. Lumileds has been able to improve on these measures; the company has developed a white light that uses yellow and red phosphors to produce a warm white light with a CRI of 93, making the light very similar to that of incandescents.
The trouble with the phosphor hack, though, is reduced efficiency. An incandescent bulb produces about 12 lumens -- a measure of light intensity -- per watt of electric power it consumes. A halogen bulb is about 15 to 17 lumens per watt. A "cool white" LED, the one that uses only a yellow phosphor, rates about 35 lumens per watt. But Lumileds' warm LED, which uses both yellow and red phosphors to produce a nicer light, produces only 22 to 25 lumens per watt -- more efficient than incandescents, but not really so overwhelming, especially when you consider that fluorescent lamps have lumen/watt measures in the 50-to-100 range.
Scott says that LED manufactures are working very hard to improve these numbers. "Today we're at 25 lumens, but we fully expect that to be well over 100 lumens per watt in time. What ends up happening when you get that high, you're not dissipating that much heat, and the possibilities become more exciting. Then we expect to see mass displacement of other markets."
How long will that take? A decade? Two decades? As with so many other things in technology, nobody can make an accurate prediction. Progress depends on many advances not yet made. LEDs are orders of magnitude more expensive than incandescent bulbs, and manufacturers will need to come up with creative ways to have them compete with standard bulbs. Eventually, though, LEDs will succeed; in time, say experts, they'll be able to produce white light better -- more cheaply and reliably -- than incandescents can, and at that time we'll probably all have to say goodbye to Edison's old light bulb.
But really, that won't be so bad. In the technology business, anything that lasts for more than a century has nothing to be ashamed of.
Get Salon in your mailbox!