Chinese semiconductor industry

Status
Not open for further replies.
G

GodRektsNoobs

Junior Member
Registered Member
Jianguo said:
There's been an EUV prototype for well over 10 years at CIOMP. This is more of an R&D project than what you would normally call a prototype associated with some sort of pre-commercial product. They've been experimenting with the components for years but there's been a lack of news on this front. The last update I heard about this was from 2017, linked below.

Please, Log in or Register to view URLs content!


There has been news involving SSMB as an EUV light source as well as EUV component testing at the SSRF. As far as I know, an SSMB EUV prototype is still in the distant future. Phase II SSMB testing is wrapping up this year with its focus on achieving highly repeatable, high power tunable EUV wavelengths. Phase III is what is supposed to be the basis for an SSMB EUV prototype for a fully integrated lithograph machine. My personal speculation is that the ongoing research at the SSRF will eventually be merged with both the SSMB and LPP EUV light source programs for dual track parallel R&D. What I know for sure is that LPP research is far ahead of SSMB R&D and is already on track for some sort of low output commercialization. This was not certain until not so long ago when rumors came out that China had a working MOPA CO2 laser in the ~30kW range. If true, then it's likely that an LPP EUV machine will be here in the next few years.
Click to expand...
It's extremely difficult to find information for precise advances in Chinese EUV systems, I think the Chinese are doing a good job keeping it under wraps. Does anyone know approximately what ASML level eqivalent was the CIOMP EUV prototype? Was it similar to SEMATECH Berkeley Microfield Exposure Tool (2003) or ASML EUV Alpha Demo Tool (2006)? If China reached Alpha Demo Tool in 2017, it would be great news since it means the Chinese EUV programme was only ~10 years behind global leading edge. ASML managed to ship the first pre-production NXE:3100 EUV only 4 years afterwards in 2010, although it was nowhere ready to mass produce chips cost-effectively. Why did it take so long to iterate from NXE:3100 to NXE:3400B before the first commercial EUV chips began mass production? It was almost 10 years.

Also, what happened to @hvpc? Haven't seen him on this forum for a while and he seem to post pretty insightful stuff.
 
tphuang

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
PopularScience said:
SMIC N+2 compare with TSMC N7+ (PPA)


View attachment 94117

View attachment 94118
Click to expand...

Alb said:
View attachment 96428
Click to expand...
escobar said:
SMIC "7nm" ("N+1") is nowhere near what some articles hyped it to be....
https://twitter.com/i/web/status/1569343390403231745
Please, Log in or Register to view URLs content!
Click to expand...

Here are some previous posts on this topic. I guess that it makes sense for some people to get really upset by the idea that China might have a 7nm equivalent. If you look at the generation 10 nm, the 89 billion/mm^2 density is basically equivalent to Intel's 10+ process and almost what Samsung had with their first 7nm process and TSMC's first 7 nm process. If you want to call it a "fake 7nm" process, that's fine, but this is enough for the server chips and AI chips that China wants to do indigenously.

More importantly, the N+2/N+2 improved process will apparently pack 127 billion and 146 billion transistors per mm^2. From what we can see in this chart, they should be ready to mass produce it next year with Kirin 9100. SMIC calls it N+2/7nm, but the density is equivalent to Samsung's 4nm process that was only achieved a year ago. And the N+2 improved process is equivalent to TMSC N5 process in density that was achieved 2 years ago and similar to N5P's power consumption that was achieved a year ago.

So if they are actually able to mass produce N+2 in first half of next year, that's a pretty big deal. And if they can produce N+2 improved by early 2024, they'd be able to get in the smartphone CPU business. It looks to me that N+2 improved would be cutting edge by today's standard.
 
olalavn

olalavn

Senior Member
Registered Member
escobar said:
SMIC "7nm" ("N + 1") không giống với những gì một số bài báo đã thổi phồng nó là ...
[MEDIA = twitter] 1569343390403231745 [/ MEDIA]
[URL opensurl = "true"]
Please, Log in or Register to view URLs content!
[/ URL]
Click to expand...
Theo như tôi biết quy trình 7nm đầu tiên của TSMC cũng là từ 14 + 14, TSMC và Huawei đã đồng tạo ra nó với DUV vào thời điểm TSMC bị EUV cấm vận ... Tôi nghĩ nếu SMIC có nhóm của Huawei từ Tân Trúc giúp đỡ ... 7nm sẽ khó thoát khỏi
 
G

gadgetcool5

Senior Member
Registered Member
escobar said:
SMIC "7nm" ("N+1") is nowhere near what some articles hyped it to be....
https://twitter.com/i/web/status/1569343390403231745
Please, Log in or Register to view URLs content!
Click to expand...
As I said in the beginning, this story was Western propaganda on the part of sanctions hawks who wanted political capital to push the CHIPS Act, so they released it right before the bill was going to a final vote. There were some people saying CHIPS was a violation of free trade and China was so far behind there was no need for industrial policy. The WSJ was giving voice to these folks because it's still a throwback to the last dying vestiges of Reagan capitalism.

Never trust a report by that D. Patel guy's firm.
 
Last edited:
T

tokenanalyst

Brigadier
Registered Member
FairAndUnbiased said:
what specific vacuum equipment for EUV? as long as they don't use diffusion pumps the cleanliness of maglev turbopumps is pretty good.

I may or may not have a tangential participation in an ASML related study on introduction of dilute hydrogen atmospheres to EUV instruments to reduce graphitic carbon buildup on optics from EUV photopolymerization, what they call "ultra clean vacuum" vs. "ultra high vacuum". That'll require a different type of pump since turbos don't have good throughput tolerance, they like as close to 0 load as possible.
Click to expand...
You are correct, but they have hold almost the same exact vacuum environment for an extended period that require a lot of sensing technology to detect leaks and others issues related to vacuum environments.
 
Status
Not open for further replies.
Top