new ultra energy efficient transistor invented

https://phys.org/news/2020-05-transistor-standard-energy-efficiency.html

It is around 10 times more efficient than a conventional transistor.
But I notice it contains some relatively rare and expensive chemical elements and that might economically limit its applications.
Also, that efficiency would probably be nothing compared to the energy efficiency of spintronic transistors when they finally come of age.

@humy said
https://phys.org/news/2020-05-transistor-standard-energy-efficiency.html

It is around 10 times more efficient than a conventional transistor.
But I notice it contains some relatively rare and expensive chemical elements and that might economically limit its applications.
Also, that efficiency would probably be nothing compared to the energy efficiency of spintronic transistors when they finally come of age.

And of Course it is a 2-D/2-D tunnel Transistor, so to actually make a circuit integrating one (or many more) of These might prove to be really difficult.

I admire what These guys achieved, jsut don't hold your breath to see it in everyday use.

@Ponderable
Well the real deal will be when spintronics gets commercial. That technology is over a thousand times more energy efficient, maybe even 100K X more efficient since it does not involve the flow of electrons to produce the transistor effect.

@sonhouse said
@Ponderable
Well the real deal will be when spintronics gets commercial. That technology is over a thousand times more energy efficient, maybe even 100K X more efficient since it does not involve the flow of electrons to produce the transistor effect.

I agree, however I don't expect that to be the case in the 20's... and I would gladly be proved wrong 😉

@Ponderable

Yep, can't wait for teraflop cell phones or petaflop laptops....

@humy said
https://phys.org/news/2020-05-transistor-standard-energy-efficiency.html

It is around 10 times more efficient than a conventional transistor.
But I notice it contains some relatively rare and expensive chemical elements and that might economically limit its applications.
Also, that efficiency would probably be nothing compared to the energy efficiency of spintronic transistors when they finally come of age.

According to the article they're using tin, tungsten and selenium. 253,000 tonnes of tin is produced annually, world reserves of tungsten are estimated at around 3.2 million tonnes with about 89,300 tonnes produced annually, and 2,000 tonnes of selenium is produced p.a. (figures from the Wikipedia pages on the various elements). It's not like germanium where only 118 tonnes or gallium where a few thousand tonnes could be extracted each year (current production is around 375 tonnes p.a.). These aren't particularly rare elements.

@deepthought said
According to the article they're using tin, tungsten and selenium. 253,000 tonnes of tin is produced annually, world reserves of tungsten are estimated at around 3.2 million tonnes with about 89,300 tonnes produced annually, and 2,000 tonnes of selenium is produced p.a. (figures from the Wikipedia pages on the various elements). It's not like germanium where only 118 to ...[text shortened]... each year (current production is around 375 tonnes p.a.). These aren't particularly rare elements.

Somehow I thought tungsten would be a rarer than that but I stand corrected.
I think it would be nice though if it was made only of the most very common elements such as carbon, silicon, iron etc.

@humy said
Somehow I thought tungsten would be a rarer than that but I stand corrected.
I think it would be nice though if it was made only of the most very common elements such as carbon, silicon, iron etc.

The one I thought might be rare was selenium, but they're all reasonably abundant. Having said that build a couple of tens of billions of processors with them and they might become rare.

@DeepThought
well, suppose for sake of argument each processor has 100 micrograms of selenium.

So 1 million processors, 100 grams. 1 billion, 10 Kg.

Don't know how close I am to the actual mass of selenium but you get the point.

Even if there was a gram then 1000 units, 1 Kg, and a million, 1000kg. a billion would get a bit dicey at one million kg but of course it is clear there would never be that much selenium in a processor.
Especially considering we are getting down to the point of counting individual atoms for transistors now, right now 7 Nm chips are coming in large production runs, 70 ANGSTROMS. like what, 20 ATOMS wide? Clearly running out of wiggle room there. I hear noises about 4 Nm chips, TEN atoms? Clearly a real problem getting down that low.

@sonhouse said
@DeepThought
well, suppose for sake of argument each processor has 100 micrograms of selenium.

So 1 million processors, 100 grams. 1 billion, 10 Kg.

Don't know how close I am to the actual mass of selenium but you get the point.

Even if there was a gram then 1000 units, 1 Kg, and a million, 1000kg. a billion would get a bit dicey at one million kg but of course it ...[text shortened]... room there. I hear noises about 4 Nm chips, TEN atoms? Clearly a real problem getting down that low.

O.k., that's the material actually in the processor, I realise you're guessing, but how much wastage is there?

@DeepThought
Sure there would be waste, my field is computer chip manufacturing and for instance one field I worked was ion implanters and that technology coats an entire wafer with whatever is implanted, the implant covers 100% of the chip and more.
So clearly there would be losses there, but still if you implant a normal level the amount per wafer is very small per wafer and I might add there are hundreds or thousands of chips on that wafer, some of which are a full 18 inches, nearly 1 full half meter in diameter. That said, the way the selenium would be used in that machine would be either some kind of a selenium containing gas coming from a lecture bottle or a sold piece in the source. What I am used to in the solid arena is stuff like pure arsenic which has a low vaporization temp, like 150 ish C or there abouts . My guess is in that tech selenium would be needed to be heated to a much higher degree than arsenic so it would probably come in a gas form.
Now coming out of the 'source' that gas goes through a magnetic ion separation and so only selenium comes out of that part of the machine. Then there is a small hole in a piece of graphite that allows a more collimated beam to go into the accelerators and focusing magnets. In each of those steps some selenium will be lost, coating onto the inner shields and onto the accelerator rings and some onto the magnets and then down the beamline spreading out some and again more lost but still the actual mass of such a process is very low, micrograms lost per run.

If the selenium is used in a sputtering tool, the 'target' would be composed of very pure selenium and in our machines it would be a slab about 125 X 350 millimeters and maybe 10 mm thick. In that case, the magnetic fields that confine the activated argon background gas, activated by some level of RF, in our case, the medical frequency of 13.56 MHz which is an industrial freebie and using up to 1500 or more watts so argon is a great bead blaster in a vacuum and the argon beats on the selenium so some selenium gets blasted off and forms a cloud where some of it gets to the platen which has the parts to be coated.
In that technology, a LOT of selenium is lost just coating the shielding plates and such and everything else in sight.
Also the magnetic field is constructed in such a way as to only have the argon ions whack the selenium target only in a shape like a horse race track.
So probably only about 20% of the selenium ever gets blasted off the target in the first place. That selenium is held in place by a solder of indium and eventually the continued use of that target would punch right through the selenium layer, hitting first the thin layer of indium and then the thick copper backing plate so that limits the lifespan of that target.
That said, if you wanted to be ergonomic you could theoretically take off the selenium, get rid of the indium and melt it down to make a new target so at least THAT would not be lost but I don't see that with our target materials, like Silicon Carbide, Tungsten/tin, SIO2, and the like, those targets just get chucked and a new one ordered.
But of course we have been known to sputter gold and you can say for sure THAT gold would be reused!