LED's Efficiency Exceeds 100% (2012) | Hacker News Reader

LED's Efficiency Exceeds 100% (2012)

phys.org

42 pointsthunderbong34 comments

Open articleView on HN

For those interested, the actual paper involved can be read. [1]

[1] https://dspace.mit.edu/bitstream/handle/1721.1/71563/Santhan...

Looks like it's also been relatively active afterward in the citations (171 total) and academic community. [2]

https://scholar.google.com/citations?view_op=view_citation&h...

"Big" surge in 2017-2019 (22, 18, 24 citations) and then fairly consistent number per year (~10) most other years.

Also, TIL you can view histograms of Google Scholar paper links. Mildly cool discovery.

I wonder if there is some underlying insight?

LEDs are both photovoltaic generators and light emitters... just not well optimized for converting light into electricity.

PV solar panels are LEDs too, just highly optimized for the opposite direction. When a panel is partially shaded (and without bypass diodes which most modern panels have) the shaded portion becomes an infrared light emitter driven by the portions in brighter light.

Based on this paper it seems something like the peltier or thermoelectric effect also exists, though obviously typical LEDs are not optimized for it and so the cooling effect / excess power is very small.

Is there some simpler unified way to describe the interactions between semiconductor electric fields/electrons, photons, and thermal energy? Maybe you could predict the best way to optimize a given material to prioritize the effect you want.

> Is there some simpler unified way to describe the interactions between semiconductor electric fields/electrons, photons, and thermal energy?

https://en.wikipedia.org/wiki/Condensed_matter_physics

Thermal energy is treated as quantum excitations called phonons.

https://en.wikipedia.org/wiki/Phonon

Oh wait you said simpler...

> Oh wait you said simpler...

Isn't that always the problem with QM? Plus there's no intuitive simplified model to explain and most analogies are so off the mark they do more harm than good. The only way to understand QM is via the math which makes it seem like hocus-pocus to most people.

The old "atoms as mini solar systems" model was easy to explain to anyone. Atom as a sun, electrons orbit around it. Yes that leaves a lot out but it is "enough" that average people can kinda understand the point of it.

Gravity and General relativity can be explained as a big rubber sheet with a heavy bowling ball in the middle. A normal person can begin to understand that spacetime has curvature. If you think about it that's something of a miracle - being able to explain something so complex so simply.

I know of nothing like that for quantum mechanics.

Simplified models and analogies give a learner a useful start of a mental model for the theory. Yes their mental model is wrong but it is useful enough. Similar to the Poverty of Stimulus principle when children learn language: learning starts with relatively shallow and sometimes inaccurate input.

Wait, where is the energy beyond the electricity coming from? Is it drawing heat from the environment, hence the references to heat pumps?

Exactly, it consumes heat.

> ... took advantage of small amounts of excess heat to emit more power than consumed.

Yep. It's taking environmental heat energy alongside electricity and converting to light.

Sundiver! Holy shit David Brin was right?

Yes, it's drawing heat from the environment: the diode itself gets cooler as it emits light, and presumably at some point there's no more available energy in this way and so efficiency goes back below 100%

Which is still amazing. An issue with high power LED is cooling. If this can be turned into a product, it solves the issue in the best possible way.

It can't for high power LEDs.

What if you point light into solar panel and use the power for LED..?

Or point the light to space for a space cooler (IIRC I've read about something like this before).

I think it is analogous to setting your tap to a teeny tiny amount of "on," and in doing so you entrain the existing droplets in your tap to fall out.

And another headline from the Dept. of Sensationally Miss-Represented Scientific Results.

Vs. reality - the ">100%" is at nano-power levels, and is powered by the LED's temperature falling.

So, while the underlying effect might have some practical application, this is ">100% efficiency" in the same sense that "turned off the LED's electric power, set it on fire, then measured the fire's light output" is ">100% efficiency".

' this is ">100% efficiency" in the same sense that "turned off the LED's electric power, set it on fire, then measured the fire's light output" is ">100% efficiency".'

It's absolutely a misleading headline. But that last part is a little harsh. The power is awfully low, but someday a partly-heat-powered LED might not be 100% useless.

DDR10 spec will have mandatory LEDs for heat dissipation reasons.

Yeah, beam that energy out of there!

This feels similar to heat pumps being >100% efficient though? Perhaps a less misleading headline would be 'Very low-power LEDs also convert heat to light'

In Brin’s Sundiver, a spaceship is able to get close to a star by dumping heat into huge freaking lasers. It would never occurred to me in 1 million years that if you have an efficient enough power source, this approach appears to work.

From 2011 - this was an interesting thing, but nothing really panned out from this since.

[2012]

Now we just need a perpetual entropy-powered photo-electric computer that uses a contained low-light LED array for internal data transfer, storage, and computation mechanisms as well as a power source. Okay, maybe not that, but this could lead to some interesting applications.

For some context, this researcher did publish quite a few more things: https://scholar.google.com/citations?user=s6Fu0woAAAAJ&hl=en

The title is click-baity, but the information is interesting.

ITT: loads of people who DNRTFA, but are here to make sarcastic jokes about perpetual motion & conservation of energy. All due to that asstastic, clickbait, lying headline.

The article isn't even a lot better. It hints at quantum fluctuations being the source, when in fact the much better-known principle of converting heat energy into electrical energy is at play (AFAICT).

TL;DR: LEDs turn out to be decent TE converters at the picowatt level. This could have applications on chips, but nowhere else seems to need this sub-nano-level kind of heat pump.

Nope. The headline is categorically false. A LED's efficiency does not exceed 100%. Science journalism is an absolute dumpster fire.

This 30 picowatt led take energy from ambiant heat. At 30 picowatts per exotic led, with a billion leds, it could extract around 30 milliwatt from the environment.

I don't understand how that means it's more efficient though. Don't you need to put that heat energy back into the environment to heat it back up again?

To me that feels like calling a magnet perpetual motion. It seems like it is, but the magnet will eventually wear out.

That's like saying this LED is over 100% efficient. It seems like it, but eventually the environmental heat around it will be gone and used up.

That in no way supports the false claim that a "LED's efficency exceeds 100%"

What if you make them nanometer thin and roll into cylinder with a solar cell (a bit like battery)

perpetuum mobile anyone?

LED produces light 230% of power put in, convert that back to electricity with PV (47.6% efficiency) to get 109.48% of electricity you put in, while conveniently cooling the surroundings.

email me to invest in my startup!

This would tend to indict an artifact in measurement rather than an indictment of the 2nd law of thermodynamics.

From the article, it seems that the extra energy is obtained from heat in the environment around the LED.

A heat pump can operate at ">100% efficiency" without breaking the laws of thermodynamics, the trick is that its moving the thermal energy around rather than merely "producing" it. It wouldn't surprise me if the LED is doing something similar.

It's a well reported effect, LEDs can emit light using electrical and heat energy. That gives better than 1 efficiency of light from electric input. Not a violation of the 2nd law, just a way to beam heat out of an environment.

Power levels are low though, so I haven't seen any reports of applications

Oh, and this may be one of the original reports, it's dated 2012.

Reply on news.ycombinator.com