« Distributed, Collaborative... Microfinance | Main | Hurricane Season, 2005 »

White Light, Less Heat

quantumdot_led_4.jpgPlease note that this article has been updated from its original text, correcting a couple of mistakes. -- Jamais

An accidental discovery at Vanderbilt University may well be the key to making light-emitting diodes the dominant lighting technology of the century. Up until very recently, the only way to make "white" light was to add yellow phosphors to bright blue LEDs. It wasn't quite right, though, as even the best "white" LED retained a blue tint. This week, we got the news that a chemistry grad student at Vanderbilt has stumbled on a way to make broad-spectrum white LEDs using quantum dots -- and in doing so, he may well have kicked off a revolution.

Michael Bowers was making quantum dots, tiny nanocrystals just a few dozen atoms across. Crystals at that scale often have unusual properties, and the ones that Bowers created were no exception. When he illuminated his batch with a laser, rather than the blue glow he expected, out came a rich white light, similar in spectrum to sunlight.

Bowers then took a polyurethane sealing liquid, mixed in some of his dots, and coated a blue LED. Although the resulting bulb -- pictured above -- is crude, it puts out white light. Its visible spectrum is similar to a typical incandescent bulb, but it puts out twice the light-per-watt, and lasts fifty times longer. One key reason for its efficiency is that it doesn't put out the infrared light typical of a regular light bulb; despite being much brighter, it's still far cooler to the touch. (The LED assembly still gets hot, however.) Completely by accident, Bowers had come up with a technology that possessed the quality of incandescent light, but none of its drawbacks.

quantumdot_led_5.jpgThe discovery was reported in the October 18 web publication of the Journal of the American Chemical Society; an abstract is available here.

These "hybrid quantum dot LED" lights should be easier and far less expensive to make than current "white" LEDs, a big step towards light-emitting diodes becoming the dominant illumination technology.

The Vanderbilt researchers are the first to report making quantum dots that spontaneously emit white light, but they aren’t the first to report using quantum dots to produce hybrid, white-light LEDs. The other reports – one by a group at the University of St. Andrews in Scotland and one by a group at Sandia National Laboratories – describe achieving this effect by adding additional compounds that interact with the tiny crystals to produce a white-light spectrum. The magic-sized quantum dots, by contrast, produce white light without any extra chemical treatment: The full spectrum emission is an intrinsic effect.
One difference between the Vanderbilt approach and the others is the process they used to make the quantum dots, Bowers observes. They use synthesis methods that take between a week and a month to complete; whereas, the Vanderbilt method takes less than an hour.
A second significant difference, according to [Bower's advisor Sandra] Rosenthal, is that it should be considerably easier to use the magic-sized quantum dots to make an “electroluminescent device” – a light source powered directly by electricity – because they can be used with a wider selection of binding compounds without affecting their emissions characteristics. [...]
The light bulb is made out of metal and glass using primarily mechanical processes. Current LEDs are made using semiconductor manufacturing techniques developed in the last 50 years. But, if the quantum dot approach pans out, it could transform lighting production into a primarily chemical process. Such a fundamental change could open up a wide range of new possibilities, such as making almost any object into a light source by coating it with luminescent paint capable of producing light in a rainbow of different shades, including white.
There are a few remaining hurdles before we start seeing white LED lightbulbs on the shelves at your local grocery store, or as commonplace tools for leapfrog lighting. LEDs are still more expensive than common incandescent and fluorescent lights, although production costs are dropping. While LEDs are very efficient, putting out more light-per-watt than incandescent bulbs, they're still not as efficient as fluorescent bulbs; the maximum efficiency possible for LEDs, however, exceeds that of fluorescents. Still, the light quality was the biggest stumbling block to broader use of LEDs, as it was an aesthetic, not a practical, deterrent.

Although replacing incandescent bulbs with white LEDs has great potential for improving energy use efficiency, I find the possibility of embedding light-emitting quantum dots into our industrial materials to be even more interesting. The breakthrough here isn't just in making it possible for LEDs to replace standard light bulbs -- although that's certainly big. The real breakthrough will come in the ways designers can start to rethink how illumination is used in our physical environment.

(Thanks to Scott Bennett and to Mike Richard at Treehugger for reminding me to post about this.)


Listed below are links to weblogs that reference White Light, Less Heat:

» White light, no heat from Rhonabwy
Very cool - a grad student stumbled through to making damn-near awesome white light with an LED. If production costs... [Read More]

» Accidental Invention - Warm White LEDs! from Treehugger
Light-emitting diodes (LEDs) are almost the perfect artificial light source: They last a long time (50,000 hours), are very shock resistant, don't produce much heat and are very energy efficient. The "almost" part is because they are still relatively... [Read More]

» White Light, No Heat from nerdshit.com
An accidental discovery at Vanderbilt University may well be the key to making light-emitting diodes the dominant lighting technology of the century. Up until very recently, the only way to make "white" light was to add yellow phosphors to bright... [Read More]

Comments (15)

Super neat. This is a timely discovery...

Alright, another step forward in the quest of life. How long before marketing to households?

One thing that just struck me is that this could be another leap in an arms race... of LED vs. fluorescent.

Fluorescent lamps have the exact same issue as the "white" LED:  they need to convert light of a short wavelength to a range of longer wavelengths.  Quantum dots appear to do this fairly well.  However, it just occurred to me that an electron in a quantum dot might be able to capture an ultraviolet photon and radiate multiple visible photons on its way back down to the ground state.  (Not being an expert in quantum physics I admit this is pure speculation, but if atoms can do it, why not quantum dots?)

This could be huge.  The major stumbling block which prevents the wholesale replacement of mercury-vapor fluorescents with xenon is that the shorter-wavelength UV photons emitted by xenon are poorly translated to visible light, especially red; to get acceptable efficiency, phosphors which generate two red photons per UV photon are required.

Maybe the researchers have been looking in the wrong place....

"While LEDs are very efficient, putting out more light-per-watt than incandescent bulbs, they're still not as efficient as fluorescent bulbs; the maximum efficiency possible for LEDs, however, exceeds that of fluorescents."

Does anyone have the current efficiency numbers for LED and fluorescent, and the maximum theorical efficiency for LEDs?

Janne Sinkkonen:

Theoretical maximum efficiency of a white light source is 200 lumens per Watt. Best incandescent bulbs are around 14-17lm/W. Fifty times the incandescent efficiency is 700-850lm/W.

The efficiency numbers claimed in the article are therefore simply absurd.

Best white LEDs are now around 40lm/W, and may soon be 60lm/W. Other _typical_ efficiencies (lm/W) from greenhouse.gov.au: low-pressure sodium 140-180 (but monochromatic orangish yellow light), HPS 50-110 (neither this is white), mercury vapour 30-50 (white enough), metal halide 72-76, fluorescent (triphosphor) 60-80.

It is also imporant to understand that the quantum dot stuff is just another fluorescent material, put on top of a short-wave LED. In a sense its efficiency is therefore limited by the efficiency of the LED.

Janne Sinkkonen:

A minor addition: the theoretical maximum efficiency for a 555nm _monochromatic_ source is 683 lumens/Watt. THis is much higher than for white because the eye is most sensitive at 555nm. Then you lose the colors, of course.

It takes a minimum of five years to go from lab to retail shelves I've heard. It's taken about 25 years for compact fluorescent lights to get advertising on TV and a "significant" market share, I'd guess less than 25% of all household lighting. (A friend of mine likes cold cathode fluorescents, though few are talking about them.)

I wonder if this is going to be a "water engine" or "Man in the White Suit" (the Alec Guinness movie) scenario - something miraculous that won't make it to the marketplace because, under the present rules, it would destroy the marketplace. There is great resistance, I've heard, to existing LED lighting because lighting manufacturers and retail outlets like selling something that has five to ten times the turnover that LEDs have. In a capitalistic economy, you like to have that cashflow. Why would you want to sell a customer a product that is basically a lifetime supply of light? The money's in the need to replace parts.

Steve Faria:

What would the lifetime usage be for an LED with the 'magic size' quantum dots incorporated? Any moves to productize this?


I have been researching LEDs lately for their role in display lighting (backlights for LCDs, and sources for projectors). These are simpler applications than general lighting, and thus will come first. They are already here, actually-- your cell phone, PDA, and digital camera have white LED backlights, made from blue LEDs that are downconverted by phosphors. These LEDs wiped out the previously-used EL backlight.

For large LCDs and projection sources (and probably general lighting too), it is far preferable to employ clusters of colored LEDs-- red, 2 greens, and a blue (to make white)-- rather than downconverting with phosphor. You get better brightness and truer color, although it comes at the expense of requiring color mixing optics.

There are several problems with LEDs:

1. Inconsistent color. Take two random LEDs and they are NOT the same color, maybe by a lot. The eye is very sensitive to this. Quite a lot of effort (and money) goes into binning LEDs so that they are color-matched.

2. Thermal management. The biggest myth about LEDs is that they are "cool" light sources. I see even Jamais repeated this urban legend. Granted, unlike incandescent lamps, LEDs do not radiate heat, so it is true that standing underneath one is not hot like it is for a big old lightbulb. But... you are not exempt from needing to get rid of the heat generated! LEDs are far from 100% efficient. So the only way to get rid of the heat is by conduction or convection. The screws of a light bulb don't get very hot because the bulb radiates its heat away, but don't touch the base of an LED! These things have to be made with pretty big heat sinks to drain off the non-radiative heat. This is a huge limitation in designing with LEDs.

Improvements in efficiency will help, of course, especially in the green. LEDs have been following something called "Haitz' Law," where the luminous efficiency doubles about every two years (ie, 10x per decade). But be aware that efficiency changes drastically with temperature, so the external conditions matter, as does, once again, the thermal management.

3. As always, cost. At present, the initial cost of installing LEDs is about 100 times that of using incandescent or fluorescent lights in a general lighting application. (The cost of *ownership* is a lot better, given the LED's longer lifetime of 40,000-80,000 hours). But it is generally agreed that costs have to come down at least an order of magnitude to even be competitive. (In display backlgihting, the factors work out differently, and LEDs only cost 2-3x over CCFL. This means penetration can just start to begin now).

Solid-state lighting is indeed a big deal, however, for multiple reasons. One reason is not always appreciated-- the color gamut. LEDs offer far more colors than the usual fluorescent or incandescent emission spectra (and phosphor response spectra). LED light can be made to resemble natural sunlight. The impact on health for people who often work inside could be noticeable, and it could help indoor plants too.

Plus all the other reasons-- no mercury, longer lifetime, etc.

Hi all -- sorry for not replying sooner, but I've been sleeping off the flu (insert Avian Flu joke here).

Janne Sinkkonen, you're quite right: the lumens/watt figure for this white LED is not 50 times that of an incandescent bulb. The lifespan is of the LED is 50 times that of incandescents, but the light output is just twice that of normal lightbulbs. That is, it's not quite as efficient as other LEDs. I'm honestly not certain whether this was an error, since corrected, in the text I was referencing, or whether I simply mis-read the item. Either way, it's now fixed.

Kim Allen, thanks for the details. I know you've been working in this area for awhile, so I'm quite sure you're right -- but it was my understanding that the quality of "white" light from combined RGB LEDs wasn't very good. That's certainly the implication from the materials linked in the post.

As you note, there are already sunlight-spectrum LED systems out there. The main advantage of quantum dot system is that it's much less costly than earlier methods, and moves us towards being able to drive down LED prices.

As for the heat, I'll update the text in the post.

If the spectral range of these quantum dots can be limited, the approaches could be combined:  use a blue LED plus phosphor to produce green up to blue, a green LED to cover green down to red, and a red LED to cover the red range.  This would give you both power efficiency and a broad spectrum with good color rendition, at a cost of greater complexity.

the science diva:

In terms of capitalism, one could look at it from an industrial standpoint. While perhaps not appealing to mass markets, if industries can use it to save energy and fuel costs it may become very attractive to corporate America, particularly in this era of rising energy costs.


No one has mentioned it specifically, but from what I can see in the picture there's another big deal brewing here. LEDs are inherently point light sources. No matter how bright we make them, they are terrible for overall, even, flat lighting. Fluorescent, gas discharge, and incandescent throw out light in all directions. This may not seem like too big a deal, but trust me, I use a lot of LEDs, and I teach others how to use LEDs, and it is a definite shortcoming at times.

From the small picture, it appears that this advance makes it possible to buld an LED based bulb that throws light in all directions - and it makes sense that it would considering how it works. I look forward to finding out how it progresses.

If efficancy of LED's are less than CFL's, why should I use it?



do you know where can ı find information about
white light hollogram


This page contains a single entry from the blog posted on October 23, 2005 12:52 PM.

The previous post in this blog was Distributed, Collaborative... Microfinance.

The next post in this blog is Hurricane Season, 2005.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.34