New LED that's over 100% efficient but doesnt break any laws

[alm]

Did you actually read the article before claiming people published in Phys. Rev. Lett. don't understand basic physics? For a theoretical background, see this (probably also behind a paywall, go shoot the publishers, not me). The short version: the entropy of incoherent light far below the lasing threshold is higher than that of the semiconductor lattice, so the emitted photons can extract up to the difference in entropy of energy (T*S) from the lattice without violating the second law of thermodynamics.

[Mechatrommer]

ok, put 15 billions of em side by side. then we'll get 1W of light from 0.43W of psu, and nice airconditioning effect

Possibly a few acres required .

no i mean using 1nm technology.

[amspire]

1nm ... Wow , that would look promising .
But then , would they be able to get down that small ? Time will tell .

I cannot see the purpose of using a small geometry. The test was done with a pretty large led, and they were emitting something like 200 million photons per second. A 1nm geometry would currently emit something like one photon per hour.

Even if they improve the efficiency a million times, a 1nm led would still be useless. You would want huge led junctions.

Richard.

[Mechatrommer]

I cannot see the purpose of using a small geometry

man! lets take practicality into calculation. say "they" are able to remove that plastic covering and mould into one (for all leds), say 3mm apart. 15bils is 122.5Kled square = 368m.sq, not too bad! what is it like? 2 football field? for 1W led? well, a PC once the size of a big room. i think we are getting there/somewhere
(ps: dont take it seriously, i'm not scientist)

[Mechatrommer]

At a decent 32nm each of 15billions leds that would mean ... 48meters ...
1nm you ask ? 1.5meters ... Wow still huge , DAFUQ ?

how do you get that? or maybe your arrangement is not optimum. i'm talking 2D array arrangement hort & vert like camera sensor. sqrt(15bils) = 122.5K led each dimension. 1nm (1e-9m) apart eqv 122.5K x 1e-9m = 0.1225mm.sq. 32nm? 3.92mm.sq! (eqv to 1 normal sized led!) but again, our Mr Richard postulates of its impossibility.

[Zero999]

Did you actually read the article before claiming people published in Phys. Rev. Lett. don't understand basic physics? For a theoretical background, see this (probably also behind a paywall, go shoot the publishers, not me). The short version: the entropy of incoherent light far below the lasing threshold is higher than that of the semiconductor lattice, so the emitted photons can extract up to the difference in entropy of energy (T*S) from the lattice without violating the second law of thermodynamics.

All I read was  the original link posted at the star of the thread and the link you've posted isn't much better.

Is there more to this?

From what I've read so far it looks like rubbish.

[alm]

The link posted by AntiProtonBoy and its references, for starters. What part of the article I referred to do you disagree with? It's not my area of expertise, but once you accept equation (2), which essentially represents the Boltzmann distribution, the rest, like eq. (2a), appears to be fairly straightforward physics.

As for the practical applications, I wouldn't hold by breath and consider this purely fundamental physics research for now.