Don't Be LED Down The Garden Path (Part 1)


Hearing a lot about "white" LEDs? So are we. If you've seen our annual Lightfair interview with the boss in Architectural Lighting, you know some of what is going on. If you haven't, that's what this article is all about.

Just so you know where we're going with this series, LEDs are great for a lot of things. Unfortunately, lighting art and artifacts isn't one of them. For low-voltage, intermittent use, monochromatic color (RGB or even Y), or long life, LEDs can't be beat. But they don't do so well where color rendition is important or UV is a problem. And wired together in banks, they loose a lot of their reported efficiency. People are making lots and lots of claims for LEDs right now. Lets take a look at LED lighting.

"White" LEDs work somewhat like fluorescent lights. Fluorescents pass current through a mercury vapor generating UV. The UV then excites phosphors on the inside of the fluorescent tube that in turn emit "white" light. "White" LEDs generate bright blue light (450nm) to excite the same kinds of phosphors. That is why the spectral power distribution above looks so strange. The spike is the LED true emission. The flat curve is the phosphor filling in some of the other colors.

UV free fluorescents are not possible. While more phosphor means better color, it also means less efficiency. Make the phosphor coating think enough to absorb all of the UV, and the outside layer stays dark. You don't get any light out.

For exactly the same reasons you can't make a "white" LEDs without a strong monochromatic color component (usually blue) overriding the phosphor emissions. You'll find a lot of variation among "white" LEDs in terms of color and efficiency as companies experiment with different phosphors, but you'll never find a true "white" LED. You can find information including color temperature and photographs of beams for dozens of manufacturers' "white" LEDs at

Scientists are experimenting with building RGB emitters into a single LED. That would create a tri-stimulus metamer to fake your eye into seeing "white" light. The problem so far is that each of the LEDs, red, green and blue, degrade at a different rate. The LEDs change color over their life. Engineers are talking about individual color monitors for each RGB LED. A "beta" version not yet on the market is supposed to have a cost "marginally higher" than high end HID fixtures. Who knows what that means?

So most of today's "white" LEDs have a strong blue component. Often the blue shows in splotches through the whiter phosphor emissions. Manufacturers mixing orange based "warm-white" LEDs with blue based "cool-white" LEDs advertise 3000°K to 6000°K color temperatures. Actual measurements are more like 3700°K to 8100°K. Mixing LEDs will help the overall average color temperature but it won't give you a more even beam. It makes things worse. Be aware that you are not going to have good color and then try to blend colors by maintaining long throws.

Beside making things look bad, blue is exactly the wrong color for preservation in a museum or archive. Blue light isn't reflected by the yellows and browns of parchments, faded textiles or ancient artifacts. It is absorbed. It doesn't aid vision. It increases damage. That is why the National Archives (NARA) set a 500nm cut off for light sources for the charter documents. Check the "white" LED power distribution again. Where would the 500nm cutoff fall?

As we mentioned above, LEDs are great for some applications. But where color is important, they just aren't there yet. You can literally get better color (and more efficiency) out of fluorescent or HID sources at a much lower cost. And, we wouldn't be NoUVIR if we didn't remind you that none of these sources meet IESNA guidelines for museum (or commercial) lighting by filtering all non-visible radiation. No UV, no IR: NoUVIR.