One would expect that reverse engineering technologies wouldn’t be impossible.
The importance of these technologies for China is huge.
Moreover, they have a massive budget and probably also own a couple of ASML’s euv lithography machines.
Why is it that they cant remake these machines?
I heard some sources say that it might take them a decade.
In: 338
My company makes parts for ASML.
Having models and dimensioned drawings are 1 step of the battle, but we have issues meeting quality and delivery requirements with all the tools available to us.
Reverse engineering something like this would net you a shittier version that’s obsolete by the time you’re done.
Shout out Asianometry on youtube. Has excellent videos covering lithography in general and quite a few on ASML’s EUV machines.
China probably does *not* own any ASML EUV machines. These are extremely high demand machines and are all called for by TSMC, Samsung, and Intel. ASML cannot build enough to satisfy demand of just these three players. They also need a fleet of PhD engineers just to run. I wouldn’t be surprised if they are 20+ years behind on this; the complexity is staggering, and likely needs a culture that promotes innovation/free thought to be successful, IMO.
Something to understand is that precision machining is HARD. People have a really hard time comprehending this. The reason why the US absolutely dominated WWII wasn’t just because of its industrial output. It was because of its ability to output precision consistently at scale.
Here’s a story. The simple ballpoint pen that people use everyday and have been around for decades. The last few decades most of them have come from China. But China was not able to produce ballpoint pens from 100% domestic components until **2017**. This is because the steel ball in the pens need to be manufactured with extreme precision and consistency or else it won’t work that well. Prior to 2017, China had to import all the steel balls from the US and the EU who were the only ones able to manufacture them with enough precision and consistency.
This is what China struggles with to this day. High tech and high precision manufacturing. The US and the EU still absolutely dominate in this field. Fundamentally it’s because doing this is hard, it’s expensive, and it requires a lot of skill/knowledge to pull off.
A more relevant story is how the CCP sank billions of dollars over the last decade to build out domestic DUV lithography machine production. They failed, they have no domestic producers of DUV lithography machines and by now the companies they started have largely gone bankrupt. They’re still completely dependent on the EU and Japan to supply DUV lithography machines.
EUV is on a completely different level from DUV. It’s like going from a Model T to a Lamborghini in the semiconductor world. ASML did not develop the machines by themselves. ASML collaborated with multiple other companies and national governments to pull it off. The US government alone invested billions of dollars into the partnerships with ASML which is why the Biden administration can put pressure on them to block exports of EUV machines to China. And the kicker is that it took well over 2 decades of R&D from the best minds in the world to develop the technology. China does not have access to that level of expertise.
If China can’t even copy DUV lithography machines, they don’t have a snowball’s chance in hell of pulling off EUV in the next decade even if they got a fully assembled machine shipped to them. Something many people don’t know is that Japan got really close to getting EUV working themselves, but the 2008 financial crisis basically forced the companies pouring money into it to stop because they couldn’t afford it. Let that sink in for a moment. An extremely wealthy country who was historically been dominant in the lithography industry and has some of the brightest minds in the field didn’t have the resources to successfully produce an EUV machine.
I’m working for one of ASML’s suppliers: ZEISS, who are producing the optical components for the lithography machines.
The main obstacle to reproducing this tech is detailed process knowledge and not just building the machine itself. Just to produce the mirrors for this system hundreds of problems have to be solved every day. This includes seemingly trivial questions like:
– What should the transport box for this mirror component look like?
– This cleaning agent seems to increase the roughness of the mirrors. Can it be replaced with something better?
– The yield has decreased by 4 %. Where in the 10-step production process did a problem sneak in?
Now multiply this by a factor of a hundred for every high-tech component in the machinery: The EUV light source by Trump or the stage system for the silicon wafers which has to be accurate to a few nanometers and withstand insane accelerations. China doesn’t just have to copy the EUV machine. And has to build a new high-tech industry and copy the world’s leading companies in laser technology, optics, piezo stage design, and more.
If you ask me they should instead attempt to build alternative approaches like electron beam lithography or ion beam lithography. Those aren’t any easier to build but at least they don’t start the race 10 years behind.
Watch [this video](https://www.youtube.com/watch?v=Fxv3JoS1uY8) showing a microprocessor in more and more detail.
Notice all those lines that cross over and under each other at about the 2 minute mark? Those are wires that connect all the components on the chip. Under neath them are the transistors and other integrated components of the chip. Every single component has connections that need to be made and wires that aren’t supposed to be connected can’t touch for the chip to work properly.
This is done in a very similar way to stone or wood block printing (you make a mask that shows where a wire should be and then remove the material that wasn’t covered by your mask). Then you do it again.
Near the very end they show a grid showing squares that are 20 nanometers across, (the chip being zoomed into is very old technology that’s much larger). That means all the wires and stuff are much, much smaller to support transistors that are 20nm rather than the size on the chip they’re zooming into.
ASML makes equipment that lets you make chips that have 3nm components meaning each small square would contain 50 modern squares.
My company makes parts for ASML.
Having models and dimensioned drawings are 1 step of the battle, but we have issues meeting quality and delivery requirements with all the tools available to us.
Reverse engineering something like this would net you a shittier version that’s obsolete by the time you’re done.
Something to understand is that precision machining is HARD. People have a really hard time comprehending this. The reason why the US absolutely dominated WWII wasn’t just because of its industrial output. It was because of its ability to output precision consistently at scale.
Here’s a story. The simple ballpoint pen that people use everyday and have been around for decades. The last few decades most of them have come from China. But China was not able to produce ballpoint pens from 100% domestic components until **2017**. This is because the steel ball in the pens need to be manufactured with extreme precision and consistency or else it won’t work that well. Prior to 2017, China had to import all the steel balls from the US and the EU who were the only ones able to manufacture them with enough precision and consistency.
This is what China struggles with to this day. High tech and high precision manufacturing. The US and the EU still absolutely dominate in this field. Fundamentally it’s because doing this is hard, it’s expensive, and it requires a lot of skill/knowledge to pull off.
A more relevant story is how the CCP sank billions of dollars over the last decade to build out domestic DUV lithography machine production. They failed, they have no domestic producers of DUV lithography machines and by now the companies they started have largely gone bankrupt. They’re still completely dependent on the EU and Japan to supply DUV lithography machines.
EUV is on a completely different level from DUV. It’s like going from a Model T to a Lamborghini in the semiconductor world. ASML did not develop the machines by themselves. ASML collaborated with multiple other companies and national governments to pull it off. The US government alone invested billions of dollars into the partnerships with ASML which is why the Biden administration can put pressure on them to block exports of EUV machines to China. And the kicker is that it took well over 2 decades of R&D from the best minds in the world to develop the technology. China does not have access to that level of expertise.
If China can’t even copy DUV lithography machines, they don’t have a snowball’s chance in hell of pulling off EUV in the next decade even if they got a fully assembled machine shipped to them. Something many people don’t know is that Japan got really close to getting EUV working themselves, but the 2008 financial crisis basically forced the companies pouring money into it to stop because they couldn’t afford it. Let that sink in for a moment. An extremely wealthy country who was historically been dominant in the lithography industry and has some of the brightest minds in the field didn’t have the resources to successfully produce an EUV machine.
This episode of Odd Lots goes into great detail as to why, if you’re interested in the answer then I highly recommend you give it a listen.
[https://pca.st/episode/607465cc-af91-446c-9944-891a51ec7e06](https://pca.st/episode/607465cc-af91-446c-9944-891a51ec7e06)
Yes the cutting edge machines they make are of incomprehensible complexity. However:
It’s not even that China don’t know or can’t figure out how the machine works. It’s that there are hundreds of thousands of parts in it, coming from literally thousands of suppliers, the supply chain is about as complex as the machine itself, and some of the parts are only made by one single manufacturer often in the US who are tightly controlled in terms of who they can export to. Every single part in the machines are bleeding edge. (I’m making this example up but:) Imagine the mirrors in it need to be flat down to the nanometer level. There are only a tiny number of companies and processes that can achieve that. Now imagine there’s some similar restriction across thousands of parts coming from thousands of suppliers. If any of the parts are imperfect then it isn’t going to do what you want it to.
This is why it would take a decade or more for China to catch up. It has taken ASML decades to get to the place they are now. Making another one of their advanced machines, even if you have one that you could strip down to the nuts and bolts, is an absolutely monumental task.
This video explains it very well: [https://www.youtube.com/watch?v=8_OOta7Y6Ik](https://www.youtube.com/watch?v=8_OOta7Y6Ik)
Basically. You need few hundred *billions* of dollars, several established corporations and a strong government to coordinate it … and you still might fail.
To be fair the progress slowed down recently so it’s easier to catch up now then it was when Japan tried it.
Latest Answers