Chinese semiconductor industry

Status
Not open for further replies.
H

hvpc

Junior Member
Registered Member
tokenanalyst said:
That is why i avoid posting posts from social media like twitter or rumors that people say, the part of the delivering of lithography machines to Chinese fabs is public knowledge that you can find in some financial documentation in the Chinese internet BUT the part were he says that SMEE is purposely fitting single wafers stages machines with dual wafer stages I haven't found it anywhere, i am reading from top to bottom. AND COURSE A SINGLE WAFER STAGES ARE GOING TO BE LOWER YIELD THAN DUAL STAGE,And specially comparing SMEE and ASML, unless you use a very powerful ligth source and very high speed stage.
Even the all powerful Nikon single stage I-line scanners have lower yield than the ASML dual stage one. 200 vs 230 wph. THAT IS JUST LOGIC.

View attachment 93795View attachment 93796

Now i have a curiosity that i can't satisfy because i can't find any reliable information, just because a post of a guy in forum.
Click to expand...
yes, Nikon has to use faster stage and higher power laser just to keep up with the dual stage system's throughput.

So, going along this line of discussion......Canon also uses single stage and their iLine and KrF throughput seems to be same or better than ASML's. That's actually pretty impressive.

For iLine and KrF I think dual stage is not a good approach. This adds extra costs to the system. Canon and Nikon able to take bigger share business from ASML in iLine and KrF because their systems are more economical and could still get the job done.

So I wonder why SMEE took the dual stage approach for their iLine and KrF. They must really like ASML's technical approach: their SSA600 looks like a copy of ASML PAS system, and now hey are building a complete line of iLine/KrF/ArF as an approximate improved copies of ASML XT dual stage platform systems.

In my opinion, SMEE should have stuck with single-stage approach to have a better tool-cost per wafer-output performance.
 
Last edited:
T

tokenanalyst

Brigadier
Registered Member
hvpc said:
is this from same section with different translation? completely different feel from the two translations. Interesting.

But the table in the blog with tool and lens NA looks right. I think I had mentioned before SMEE's best advertised SSA600 90nm system is quite a big step away from immersion. From their papers, it's a 0.75NA system so not even the same as the modern Nikon/ASML ArF system that has 0.93NA. So on the table, we see that modern KrF systems from Canon and ASML could hit same 90nm resolution as SMEE SSA600 ArF system. So, SSA600 not as advanced or equivalent to current ArF dry system as many people think it is.

So some info correct and some incorrect. Not so atypical for blogs or articles on the web.
Click to expand...
To be fair the paper is from 2011 when they did almost everything by themselves with some Chinese universities, for some time SMEE itself looks like didn't wanted to deal with front end scanners preferring to focus in packaging,FPD and power devices, that is until the government step up with the project 02 to develop a back them a 65nm immersion scanner, U-Precision was founded in 2014, RSlaser 2016 and Guaowang optical in 2020 by E-Town.



SMEE_2011.png


And for the Military and China space program that scanner has must had been a hit.
 
Last edited:
H

hvpc

Junior Member
Registered Member
tokenanalyst said:
To be fair the paper is from 2011 when they did almost everything by themselves with some Chinese universities, for some time SMEE itself looks like didn't wanted to deal front end scanners preferring to focus in packaging,FPD and power devices, that is until the government step up with the project 02 to develop a back them a 65nm immersion scanner, U-Precision was founded in 2014, RSlaser 2016 and Guaowang optical in 2020 by E-Town.



View attachment 93801


And for the Military and China space program that scanner has must had been a hit.
Click to expand...
This is basically the SSA600. In 2011, this is the SSA600/10. As you can see, the imaging resolution of this 2011 version is 130nm.

There are more papers after the one you posted with same general spec. But I think SMEE added more capabilities since I saw some of their paper on lens wavefront abberration correction and lens heating correction, which presumably were incorporated into their latest advertised SSA600/20 with 90nm resolution. The SSA600/20 also has faster throughput than the 50wph of this older system.

tokenanalyst said:
That means working in very high speed wafer stages and very powerful lasers which is a weakness right now instead of dealing with the control algorithms of a dual wafer stage.
Click to expand...
True. True. From technical side, it's a decision picking the lesser of two evil. Put in work to build better laser or deal with difficulty to control two stages at fast speed (if you recall from their IPO filing, beside controlling the two stages, U-Precision's dual stage speed still needs to be made faster to match ASML).

From business side, both approach has technical challenges so that cancel each out, but the single stage would be cheaper and made better sense. This is why I'd lean towards good enough tech and lower price tag. I don't think fabs willing spend premium for less critical technology like iLine & KrF. I'm biased on this one, since I believe ASML's iLine/KrF dual stage approach is flawed because of high cost lead to them unable to have clear advantage over Canon/Nikon.
 
Last edited:
T

tokenanalyst

Brigadier
Registered Member
hvpc said:
This is basically the SSA600. In 2011, this is the SSA600/10. As you can see, the imaging resolution of this 2011 version is 130nm.

There are more papers after the one you posted with same general spec. But I think SMEE added more capabilities since I saw some of their paper on lens wavefront abberration correction and lens heating correction, which presumably were incorporated into their latest advertised SSA600/20 with 90nm resolution. The SSA600/20 also has faster throughput than the 50wph of this older system.
Click to expand...
I think i posted that one here, i think is from 2013. Following their patent activity looks like they put a bigger effort to products that hey where making money from, packaging, FPD and power devices, even they manage a deal with ASML to help them with that in 2017 when the technologies for the project 02 passed acceptance, until the tensions between China and the US started heat up, they shifted their focus a least publicly to the "project 02" and you can see shift in patents from then to today.
 
H

hvpc

Junior Member
Registered Member
tokenanalyst said:
I think i posted that one here, i think is from 2013. Following their patent activity looks like they put a bigger effort to products that hey where making money from, packaging, FPD and power devices, even they manage a deal with ASML to help them with that in 2017 when the technologies for the project 02 passed acceptance, until the tensions between China and the US started heat up, they shifted their focus a least publicly to the "project 02" and you can see shift in patents from then to today.
Click to expand...
Could you elaborate a bit more on your "(SMEE) shift in patent for then to today" observation?

By the way, it just hit me, iLine doesn't use laser. KrF doesn't need high power laser, RS Laser's KrF offering is plenty capable to handle fast single stage. So, one less reason against leaning to single-stage iLine & KrF. :D
 
A

antiterror13

Brigadier
tonyget said:
As far as I know，SMEE isn't the sole organization in China which is responsible for lithography machine development. There is CETC 35th bureau also developing lithography machine
Click to expand...

Is CETC only for military ? as far as I know Chinese military chips and lithography very much 100% self sufficient, I think with 65nm or 45 nm lithography which is very much enough for military
 
Status
Not open for further replies.
Top