thermoionic cooling and thermotunnel cooling?

[coppercone2]

has anyone here worked with this kind of stuff?

[Sceadwian]

Looks like it has little to no practical usage possibilities, certainly not for electronics.
https://en.wikipedia.org/wiki/Thermotunnel_cooling

[coppercone2]

the biggest problem is the 10nm flatness requirement, which is fulfilled by using piezeo actuators and making mated non calibrated surfaces using electroforming

You can approach the flatness with lapping, but the requirement is something like 300 cesium atoms separation. I assume they make something as flat as possible and then use electroforming.

You are right that thermoionic cooling has little use for conventional electronics, but thermotunnel can work at room temperature, if you can set the thing up right. The whole idea is that it can work at room temp.

[Sceadwian]

To what effect? I've got no sense of this, would it be worth it? Peltier's were supposed to be the cat's pyjama's and not that they don't have their uses but they're a niche product.

the biggest problem is the 10nm flatness requirement, which is fulfilled by using piezeo actuators and making mated non calibrated surfaces using electroforming

You can approach the flatness with lapping, but the requirement is something like 300 cesium atoms separation. I assume they make something as flat as possible and then use electroforming.

You are right that thermoionic cooling has little use for conventional electronics, but thermotunnel can work at room temperature, if you can set the thing up right. The whole idea is that it can work at room temp.

[coppercone2]

It would be like 100 watt/cm squared cooling or better with a vacuum barrier between the hot and cold sides, so you would only have leakage radiation to reheat the cooled side.

With a peltier you have the combined area of all the semiconductor junctions to act as a thermal shunt.

I am not sure how carnot efficiency is related to the thermal leakage, or if its related at all, but it would be better.

[coppercone2]

I wonder if robrenz knows how to make mating 10nm spaced surfaces that are large...

[Kleinstein]

It was a idea to make something similar to a Pelletier cooler.  The main idea is to avoid the thermal conduction from contact and still get electric current across by tunneling.  The 1st problem is to get the short distance over a large area. The 2 nd less known problem is that with such a small gap, there will be near field electromagnetic coupling, which tends to increase the radiation coupling. 

So in short it does not really work.

[coppercone2]

How does the EM field increase thermal coupling ?
Can you explain the effect?

And it has been done before, but it looks like the Georgian paper is a complete pain in the ass to execute. Looks like you need heavy experience with film deposition and semiconductor like processes to copy their work.

[Marco]

Thermotunneling needs to put surfaces so close together that it's almost impossible to isolate the surfaces sufficiently and probably requires UHV.

Thermionic cooling requires extra energy to free the electrons at room temperature, either an accelerating grid or photons ... either way I'm not sure you can regain the energy expended to accelerate the electrons to emission, maybe with a grid if you can manage to minimize grid losses. Still the scale you have to work with is more reasonable than tunneling and even with the losses it might be interesting for on-chip cooling.

A hobby for university students with access to a MEMS production facility.