D
Deleted member 15949
Guest
Your assuming that at high-NA EUV, the frontier isn't closed
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
I recently found out this video was referring to a materials breakthrough based on platinum sulfide. I found the following links related to this video.
They are saying that they created a new graphene like material to eventually replace silicon. When they said graphene like, they are basically saying it is a 2D material, meaning it is atoms thick and allows electrons to flow faster and more directionally through the material, unlike silicon which has a diamond like tetrahedral structure that allows electrons to move in all directions, which makes electron flow relatively inefficient, slower and prone to thermal issues. This is why graphene was always hyped up as a low power, high performance silicon replacement. The problem with graphene is that it isn't a semiconductor material, which means it doesn't have a bandgap. You need the bandgap in order to be able to turn transistors off, otherwise you can't make logic circuits.
This platinum sulfide semiconductor has a "wide and adjustable" bandgap which is probably "adjusted" by how many atoms thick the 2D structure is. They compared their platinum sulfide semiconductor with what TSMC has already researched with bismuth semiconductors and claim they can use it to eventually fab processes down to 1nm without quantum tunneling.
This is all good, but the way the reporters presented this, they make it sound like this is just around the corner when it's still in research. What I find more interesting is the public way this materials breakthrough is presented. If this was so important, why would they publicly report this and present it as a breakthrough for Humanity when this would simply be used by the American semiconductor alliance to stay ahead of China? This reminds me of when the China 2025 goals were publicly announced. It sounded good on the surface but it was actually the final impetus needed for China's enemies to fully zero in and target China as their number #1 target to destroy. That was stupid of China and so is this. Anything secret should stay secret. Only when the cat is already out of the bag should such secrets be allowed to be spoken.
What's the reason though? The theory is that you can make everything thinner. But you already waste a ton of vertical headspace as is, which is why 3D packaging architecture like chiplets, X-tacking and through silicon vias are getting popular.
The other reason is electron mobility. Good reason, but then there's GaAs and SiGe.
The killer is lack of inert native oxide and nitride. Without a native oxide it is much harder to deposit an epitaxial insulator and without inert native nitride you can't deposit an inert epitaxial blocking layer. That's why even for ultra high performance lab chips with exotic materials at most they still usually use something like SiGe or SiC.
I understand why people try to get 2D materials for logic, and this is worthy research, but it is not ready for near term by far because the performance requirements are nowhere near where they'd be justified by the vastly high cost and lower yield.
There's lots of other applications for 2D materials such as battery tech, sensors, RF, MEMS, heating elements, etc. Why try to shoehorn it into logic?
They sandwiched graphene between boron nitride which was the blocking layer. This was in a landmark paper about transistor use with graphene published in Science Magazine. You will find it under the title "Field-effect tunneling transistor based on vertical graphene heterostructures". Not saying this is the way to go, this is all research.
In the case of other 2D material candidates, not saying they are ready or better than existing pathways like chiplets, Xtacking, nanosheets, etc. Those are clearly ready or almost ready. However, beyond 2nm, unless thermal issues can be definitively solved for various layering techniques, we're going to hit a wall in 10 odd years. When that time arrives, silicon will have to be replaced with something else. Imo, it's going to be some variation of a 2D semiconductor, perhaps platinum sulfide as recently discussed.
A complete chip blockade from the China market is not a serious talking point.
They should do it, if they think they can get away with it.
That is a quick way to commit economic and technology suicide.
I wouldn't say existential.
My guestimate is that China could cover most of its domestic semiconductor consumption with current fab capacity inside China.
So it would be a tough few years until replacement fabs are up and running, but China will get there.
Yes, but that requires deposition of a boron nitride layer which means introducing expensive precursors such as borane, which are toxic, flammable and particle forming in presence of water which can destroy pumps and contaminate other wafers as boron is a dopant if anything goes wrong. There might be lattice mismatch or poor adhesion.
Meanwhile for silicon you only need to introduce pure gases which are not as flammable, toxic or particle forming such as water vapor to form the oxide layer. You can form them to arbitrary thickness. And they come as epitaxial films.
I actually think electronic logic is just going to slow down and hit the brick wall until quantum computing. Each new process gets exponentially more expensive for less and less gain which means longer and longer time on market to recoup costs at higher end prices for customers who see little difference in new products.
- Status