China Aims to Use Particle Accelerators to Build Chips and Evade EUV Sanctions(tomshardware.com)
tomshardware.com
China Aims to Use Particle Accelerators to Build Chips and Evade EUV Sanctions
https://www.tomshardware.com/news/china-aims-to-use-particle-accelerator-to-build-chips-and-evade-euv-sanctions
13 comments
Doubtful.
ASML didn't choose to laser vaporize rapidly moving miniscule tin droplets because it was easy. It was simply easier than making a synchrotron work.
ASML didn't choose to laser vaporize rapidly moving miniscule tin droplets because it was easy. It was simply easier than making a synchrotron work.
I think this is a bad argument. It's like saying only because someone clever choose path A it is wrong to choose path B.
Free electron lasers are not without their difficulties either, so I guess China assessed how well they could follow ASML and, given their difficulties, decided to learn from someone else's mistakes instead of making them themselves.
They still might learn from their own mistakes, but at least they are trying something different. I'm not sure vaporizing tin-droplets can be improved enough to solve all of ASML's problems.
They still might learn from their own mistakes, but at least they are trying something different. I'm not sure vaporizing tin-droplets can be improved enough to solve all of ASML's problems.
IIRC, they chose it because it's more compact and easier to install in a factory than a synchrotron, and also are a continuation of their current product evolution. I don't think they'd be able to make a synchrotron suitable for this that could be shipped in a standard container (or loaded onto a single truck).
There were also issues getting the light sufficiently collated from a synchrotron.
This Asianometry talks about the engineering: https://www.youtube.com/watch?v=5Ge2RcvDlgw
Now, China may decide that it doesn't need throughput and that will definitely help. Having a facility that can simply do it even if it can't do it at speeds that a commercial fab would accept would be fine for strategic reasons.
However, it will be 20+ years of engineering.
And that's just the light source.
The optics, for example, are a whole other engineering project.
This Asianometry talks about the engineering: https://www.youtube.com/watch?v=5Ge2RcvDlgw
Now, China may decide that it doesn't need throughput and that will definitely help. Having a facility that can simply do it even if it can't do it at speeds that a commercial fab would accept would be fine for strategic reasons.
However, it will be 20+ years of engineering.
And that's just the light source.
The optics, for example, are a whole other engineering project.
IIRC EU/EUCLIDES exploration of EUV light source concluded synchrotron easier/cheaper to _technically_ operate long term, but exploration limited to 8W, compact designs. As far as I can tell, getting to higher power = scaling up size to 100s of m = made it logistically infeasible to build/plan fabs around. The Tsinghua SSMB design is 40m diameter, which is compact enough to be useful. Also ASML chose laser plasma because they thought synchrotron was too expensive to build, but years of technical challenges and a machine costs 200m, next gen will be 400m... that's already the price of multiple small synchrotrons. TLDR is laser plasma, works, is very expensive to operate, and ASML is probably locked in due to sunk cost.
What I really like about this is that it's a completely different approach from what the other big player is doing. Evolution in nature is based on small iterative variations to locate local maxima without ever regressing to a worse fit on its way to an even better maximum. Very few times we see someone doing something wild in hopes they'll end with something better in the end.
Like this? https://www.youtube.com/watch?v=0igQuerc3J0
The article says its SSMB which is a hybrid of synchrotron and a linear free electron laser. https://www.quora.com/What-is-a-Synchrotron-SSMB-generated-E...
This remembers me USSR, when they hit by sanctions, they also try to create some extraordinary solution, centralized way, because of limited total resources, but better concentrated than in free market economy.
Sometimes, centralized way could even get ahead decentralized free market way. Will see.
Sometimes, centralized way could even get ahead decentralized free market way. Will see.
This is just an alternate EUV light source. I don't think the light source is the hard part of building systems for 2 nm lithography.
ASML does that by vaporizing tiny droplets of tin while they are in free-fall in a vacuum by hitting them with very precisely timed laser pulses.
This light source took years to develop and ASML opted to do it this way because the development risks were lower considering their previous experience with similar less extreme UV light sources.
This light source took years to develop and ASML opted to do it this way because the development risks were lower considering their previous experience with similar less extreme UV light sources.
https://en.m.wikipedia.org/wiki/Diamond_Light_Source
I notice a major £500m upgrade was recently announced:
https://www.diamond.ac.uk/Home/News/LatestNews/2023/060923.h...