230% Efficient LEDs

[monkeyboy]

So if we run 100 billion XP-G's at 30 picowatts a piece, we'd be getting 6.9W of light for 3W of electricity. I think. That many zeros messes with my brain.

Interesting idea. Back-of-the-envelope calculation:

You'd need an area of 1.5 km x 1.5 km and more money than Bill Gates. Not sure you'd even be able to see a dim glow with 6.9W spread over that area. Imagine the cabling you would need and the power losses through that length of cabling!

How about this for an idea: put the die in contact with an efficient photovoltaic solar cell which simultaneously powers the die and extracts energy from it, and you have a machine that extracts heat from the environment and directly produces electricity. I'm sure that violates every law of thermodynamics though.

 

[mvyrmnd]

How about this for an idea: put the die in contact with an efficient photovoltaic solar cell which simultaneously powers the die and extracts energy from it, and you have a machine that extracts heat from the environment and directly produces electricity. I'm sure that violates every law of thermodynamics though.

It would only break the rules if it ran forever. With no external input of power after the initial power up, it would simply fade out over time.

 

[uk_caver]

You'd need an area of 1.5 km x 1.5 km and more money than Bill Gates. Not sure you'd even be able to see a dim glow with 6.9W spread over that area. Imagine the cabling you would need and the power losses through that length of cabling!

If starlight is ~10^-4 lux(10^-4lm/m^2), with white light at ~350lm/W the 6.9W of light would be 2400lm, spread over the 1.5x1.5km, about 10x as bright as starlight, if I haven't screwed up calculations.

 

[AnAppleSnail]

I'm not sure you can get useful light at any wavelength with an input power of 69 picowatts. Pragmatism aside, as far as I know an LED will *always* produce its intended color of light as long as the drive voltage is within spec, regardless of how tiny the amperage may be.

They say they have an LED with a "very small band gap." To me that means low-energy photons. ie, they have produced an infrared LED emitting 'bonus' infrared light.

 

[uk_caver]

If so, when they say 'these initial results provide too little light for most applications', I wonder what applications there would be for a very dim infrared LED?

 

[jtr1962]

My take on all this isn't that the LEDs in the experiment have any practical applications, but rather that the experiment helps to understand where the losses occur. This in turn could eventually lead to LEDs which are nearly 100% efficient. Once you get past 75% or 80% efficiency, small gains may not save you much power, but they could greatly decrease heat-sinking requirements. Going from 80% to 90% cuts your heat-sinking requirements roughly in half. Going to 95% cuts them in half again.

 

[AnAppleSnail]

If so, when they say 'these initial results provide too little light for most applications', I wonder what applications there would be for a very dim infrared LED?

Either removing a few picowatts from something inefficiently, or it's only mentioned because the science journalists misinterpreted things, or else I'm misreading what the band gap means.

 

[fyrstormer]

If so, when they say 'these initial results provide too little light for most applications', I wonder what applications there would be for a very dim infrared LED?

One application that comes to mind is a cooling surface for cryogenic experiments. The crystalline surface of the emitter would help the cryogenic substance radiate away its remaining heat.

 

[ssvqwnp]

I think you guys are looking at this the wrong way... We could cover our existing LED heatsinks with these LED and let them run off the heat, that way we can lave LED-powered LED!



Yo dawg, I heard you like LED, so we put LED on your LED so your D can EL while your D EL's... (I'm new at this, forgive me)

Someone had to say it.

 

[ma_sha1]

...........

 

[TEEJ]

If starlight is ~10^-4 lux(10^-4lm/m^2), with white light at ~350lm/W the 6.9W of light would be 2400lm, spread over the 1.5x1.5km, about 10x as bright as starlight, if I haven't screwed up calculations.

You are assuming that ALL of the stars combined in the night sky are covering the ground with a given Lux

 

[uk_caver]

You are assuming that ALL of the stars combined in the night sky are covering the ground with a given Lux

Well, I was using generally given figures for illumination by starlight (presumably on a fully clear moonless night away from any artificial light), and I was assuming the figures I was progressing from were correct.
As far as the LEDs were concerned, the assumptions I was making included:
a) the output being visible light
b) a rough watt - lumen/lux conversion of about half the green optimum
c) a basically infinite illuminator, such as extensive LED-covered panels slung overhead much wider than they were high above me, such that the intensity of light hitting the ground was the same as the intensity of light leaving the emitting panels (so definitely nothing like a flashlight actually lighting an area much larger than its emitter to starlight-like levels).

The first one of which may be wrong, the second inaccurate, and the third impractical (though putting some kind of upper bound on the possible intensities of illumination)

Though it was only an estimate to get some rough handle on the kind of light/energy levels involved.

 

[Mattaus]

Yo dawg, I heard you like LED, so we put LED on your LED so your D can EL while your D EL's... (I'm new at this, forgive me)

Someone had to say it.

Lol, don't worry ssvqwnp, I get what you tried to do there. Pretty good effort IMO.

 

[pretmetled]

Yeah, ran into that article earlier on. Nice heat pump.

 

[Lynx_Arc]

There is something wrong with the article. You don't measure light in watts at all and measurements for light have nothing to do with power itself so trying to say picowatts of light is meaningless because all they have done is proved that they can make more efficient light than thought before. They must have started with some sort of "idea" that so many watts = so much light which would indicate if they were correct in the first place that number would not be unable to be surpassed. In other words their experiment only proves that their "idea" of "watt of light" was incorrectly too low to begin with and they doubled the amount which to me means they were off 231% in their initial calcuations of "watt of light" to begin with.

 

[AnAppleSnail]

They must have started with some sort of "idea" that so many watts = so much light

1 lumen = "1 lm = 1 cd*sr[URL="http://en.wikipedia.org/wiki/Steradian"]"[/url] and 1 cd is "the luminous intensity, in a given direction, of a source that emits monochromatic radiation of frequency 540×1012​ hertz and that has a radiant intensity in that direction of 1​⁄683​ watt per steradian."

Light, like sound, can be expressed as an amount of energy. Doing so accurately may be harder than I thought, but it is done.

 

[Lynx_Arc]

1 lumen = "1 lm = 1 cd*sr[URL="http://en.wikipedia.org/wiki/Steradian"]"[/url] and 1 cd is "the luminous intensity, in a given direction, of a source that emits monochromatic radiation of frequency 540×1012​ hertz and that has a radiant intensity in that direction of 1​⁄683​ watt per steradian."

Light, like sound, can be expressed as an amount of energy. Doing so accurately may be harder than I thought, but it is done.

I still stand with the notion you cannot get energy from nothing. There is no such thing as 230% efficiency unless you have a baseline to start with that is not truly 100% efficient. The laws of thermodynamics do not include energy being created from nothing at all which is what this "discovery" to me seems to be concluding innaccurately. I somewhat doubt they are using this formula to get their energy (in watts) from but instead basing it upon a standard of light output (in watts) based upon a known not 100% efficient energy source and comparing it with that. If they were using a 40% efficient light source then achieving nearly 100% efficiency then they would achieve this 230% figure.

 

[pretmetled]

There is something wrong with the article. You don't measure light in watts at all and measurements for light have nothing to do with power itself so trying to say picowatts of light is meaningless ...

Wut?

I guess that all those laser output specifications in Watt are meaningless then.

Translation: measuring light output (you know, photons schmotons) in Watt is perfectly valid.

 

[Lynx_Arc]

Wut?

I guess that all those laser output specifications in Watt are meaningless then.

Translation: measuring light output (you know, photons schmotons) in Watt is perfectly valid.

I don't see it being done by LED manufacturers.... it is lumens/watt..... not just watts. I still think this 230% is bogus because if something is totally efficient it is 100% at best anything over that means (via the laws of thermodynamics) that you must have energy being introduced from some unknown source (miracle).

 

[pretmetled]

I don't see it being done by LED manufacturers.... it is lumens/watt..... not just watts. I still think this 230% is bogus because if something is totally efficient it is 100% at best anything over that means (via the laws of thermodynamics) that you must have energy being introduced from some unknown source (miracle).

Random google hit du jour:

http://www.cree.com/products/pdf/XLampXT-E_ROY.pdf

Page 6, radiant flux specified in mW.

As far as I'm concerned the penchant for using lumens has everything to do with convention in a particular business (lighting) and nothing with it's merit over SI base units (as opposed to derived units).

Case in point: I've noticed that for royal blue leds it's quite customary to specifify radiant flux in Watt. And for regular boring blue leds that are the same in every respect (except for a few nm in wavelength ) then suddenly lumens is all the rage.

Anyways, the 230% claim is rather arbitrary if you ask me since it's as open loop as you can get. But thermodynamically speaking there's nothing magical going on there. It's analogous to pumping up hot water and then stuffing it through a generator. Suppose the pump etc is 100 Watt and you can load the generator output by 230 Watt then your clever marketing monkeys can say "tada! 230% efficient! ^_^" in exactly the same way as these guys are doing.