Chinese semiconductor industry

Status
Not open for further replies.

tokenanalyst

Brigadier
Registered Member
Everything about Extreme Ultraviolet lithography is Extreme. Errors are measured in picometers rather than nanometers. Meter diameter intercalated multilayer mirrors flat down to the dozen of picometers, light wavelength aberration down to the dozen of picometers, also the wafer stage has to be more precise to take advantage of shorter wavelength and also has to be maglev. You have to hit a molten drop of tin dozen thousands of times per second with a laser, extreme vacuum requirements, extreme anti-vibration requirements, really complex hydrogen tooling to clean the mirrors, the energy waste is huge like a megawatt to produce a few hundred of watts of EUV light, very accurate EUV sensors and so on.
LPP require complex MOPA lasers, DPP requires complex electric tooling and cleaning, SSMB and Synchrotron is basically a particle accelerator.

Photos of the machine itself looks complex due the huge quantity of cabling used to connect sensors, power units and others.

But is also is a Extreme mistake to think that given enough extreme pressure and extreme market forces, companies supported by an entire country with big resources cannot develop and commercialize EUVL technologies their own. And more if such country ALREADY have develop over the past years relevant EUV technologies on its own.
 

latenlazy

Brigadier
Everything about Extreme Ultraviolet lithography is Extreme. Errors are measured in picometers rather than nanometers. Meter diameter intercalated multilayer mirrors flat down to the dozen of picometers, light wavelength aberration down to the dozen of picometers, also the wafer stage has to be more precise to take advantage of shorter wavelength and also has to be maglev. You have to hit a molten drop of tin dozen thousands of times per second with a laser, extreme vacuum requirements, extreme anti-vibration requirements, really complex hydrogen tooling to clean the mirrors, the energy waste is huge like a megawatt to produce a few hundred of watts of EUV light, very accurate EUV sensors and so on.
LPP require complex MOPA lasers, DPP requires complex electric tooling and cleaning, SSMB and Synchrotron is basically a particle accelerator.

Photos of the machine itself looks complex due the huge quantity of cabling used to connect sensors, power units and others.

But is also is a Extreme mistake to think that given enough extreme pressure and extreme market forces, companies supported by an entire country with big resources cannot develop and commercialize EUVL technologies their own. And more if such country ALREADY have develop over the past years relevant EUV technologies on its own.
Tbh the thousands of tiny drops a second with a laser problem sounds pretty impressive and complicated but it’s actually one of the mechanically simpler dynamics in the machine. Periodic coordination is extremely easy with digital controllers but was for the most part a trivial well understood mechanism even when everything was controlled by gearboxes. Vibration control and positioning are harder problems. Optimal excitation dynamics for the tin plasma is harder. Debris control and collection is harder. So yeah…the physics involved can be called extreme but that doesn’t mean every dynamic is complex. Usually in engineering the most difficult to deal with problems aren’t the most impressive looking ones.
 

latenlazy

Brigadier
Tbh the thousands of tiny drops a second with a laser problem sounds pretty impressive and complicated but it’s actually one of the mechanically simpler dynamics in the machine. Periodic coordination is extremely easy with digital controllers but was for the most part a trivial well understood mechanism even when everything was controlled by gearboxes. Vibration control and positioning are harder problems. Optimal excitation dynamics for the tin plasma is harder. Debris control and collection is harder. So yeah…the physics involved can be called extreme but that doesn’t mean every dynamic is complex. Usually in engineering the most difficult to deal with problems aren’t the most impressive looking ones.
(Though in retrospect I probably should have said electronic controllers because both digital and analog electronics makes this stuff very straightforward)
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
Geopolitical implication of China's investment into mature fabs, especially for auto chips. There is just less demand left for everyone else. Taiwan can keep trotting TSMC out there to score favors from Western gov't. What's the justification for these fabs to happen if those European chipmakers are having trouble selling their production?

As I said before, there is real geopolitical implication for China to keep advancing its own chipmaking industry since TSMC's end customers are not in Taiwan, but other countries. China represents half of world's chip market. If Chinese chipmakers can supply their own market, what does TSMC's future look like? Of course, we are still at least 5 years from that point, but TSMC is basically Chinese EUV away from losing its position in the world.
 

FairAndUnbiased

Brigadier
Registered Member
Most of modern technologies would be black magic for your average normie and 99% of journalists are the same normies just able to write eloquent bs for living.
forget modern technologies... even the steam engine is too complicated for most people to understand.

"this alien black magic engine can use ANY combustible fuel and can convert expansion work derived from combustion into rotary motion! wow! it is sooo difficult to make though, as it requires corrosion resistant, high temperature yet cheap alloys, precision gearbox machining and an automatic control system. Based on the complexities I can only guess that it will be millennia before China masters this alien tech."
 
Status
Not open for further replies.
Top