Never underestimate your adversary. We can score a similar victory with Carbon nano tubes. Lete run TSMC out of business.
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
Never underestimate your adversary. We can score a similar victory with Carbon nano tubes. Lete run TSMC out of business.
The Americans shouldn't be coercing TSMC as they are. That said, TSMC's ties to the US go deep.
Founder Morris Chang
- Born in Zhejiang, fled in 1948
- Bachelor & Master of Mechanical Engineering at MIT
- Promoted to Manager of Engineering Department at Texas Instruments in just 3 years
- Stanford PhD, funded by Texas Instruments
- 25 year career at Texas Instruments, rising to Head of the Semiconductor Division
- All of this before founding TSMC in 1987
In a way you could say this guy was "made in America" and TSMC was "made in America". One reason I always say the idea that America is China's implacable enemy is completely false. It was a political choice not to prioritize good relations.
I agree with you. We need to have a working relationship with the US. I think that could be possible with a new Administration. Whether America will continue to try to undermine us. I definitely think they will, but that's the perenial problem of course.
Ignore my previous comments about SMEE 28 NM machine being able to make 7NM chips. I've gotten someone who can read Chinese to translate the SMEE website for me. It seems that SMEE uses the definition of resolution as the smallest node that the machine can produce. This is different from how ASML defines resolution.
ASML machine with 38 nm resolution can produce 7nm chips.
SMEE 28nm resolution cannot produce 7nm chips. The smallest node/chip that SMEE 28 nm can produce is 28nm.
So global times is right and I am wrong. This means that this isn't the breakthrough we are looking for and we are literally many years away from domestic lithograph that can produce 7nm chips.
ASML machine with 38 nm resolution can produce 7nm chips.
SMEE 28nm resolution cannot produce 7nm chips. The smallest node/chip that SMEE 28 nm can produce is 28nm.
So global times is right and I am wrong. This means that this isn't the breakthrough we are looking for and we are literally many years away from domestic lithograph that can produce 7nm chips.
Last edited:
Skywatcher
Captain
That doesn't take into account multiple patterning.
I'm pretty sure 28nm is with multiple patterning. Without multiple, the hard limit is more like 56nm, while their old machine is 180 nm.
Skywatcher
Captain
That's probably questionable, since since SSC600/10, with a 0.75 NA, which uses a 193nm KrF power source, is would presumably be 180/193nm in single exposure, 90nm with multiple exposures.
However, immersion lithography would improve single exposure for a lithographic machine with an least 0.9-1 NA to at least 40-45nm* (it's highly unlikely that Beijing Guowang Optical Technology**, which invested $1 billion RMB in 2019 in capital investment and is reputed to supply the optical system for the SSA/SSB800, is still stuck on 0.75).
*
**
Last edited:
This is going with the assumption that ssa800 is using immersion? Even if we assume it does and also has a NA of 0.9-1. Single exposure is only 40-45 nm. Practical single exposure is probably lower than that. So if you half it, its about 28nm after double exposure......
Skywatcher
Captain
Immersion technology has been around for the last 15 years, it'd be more suprising if SMEE didn't utilize it
And 40-45nm is still pretty close to 38nm (we're assuming that the NA does not exceed 1.0, a larger NA will increase practical single exposure performance). At this point, to get to 7nm, we have to look at the overlay performance.
Hmm ok, that makes more sense. ASML quotes their machines at 38nm resolution, which is single exposure resolution. If we assume that ssa800 can achieve a single exposure of around 38 or 40nm (depending on Numerical aperture and wet immersion), Theoretically assuming good overlay performance, 7nm would be possible?
- Status