用户6878647890: 这不扯淡吗,icrd都被制裁了,不用纯国产线想生产都不可能
havok: 人家ICRD早有准备,有一条可延伸至14nm的小产线把原本计划进口的设备全换成国产。这条产线也是为验证28nm机准备的
havok: 此国产验证线投资20多亿,规划有3台光刻设备和涂胶显影设备,总设备数小于100台,其实就是嘉定产线的微缩版。设备验证完后就能拉到大产线去。
User 6878647890: Isn't this nonsense?icrd has been sanctioned, it is impossible to produce without pure domestic production line
havok: ICRD has been prepared for a long time, and has a small production line that can be extended to 14nm to replace all the equipment that was originally planned to be imported with domestic production. This production line is also prepared for the verification of 28nm machines
havok: This domestic verification line has an investment of more than 2 billion. It is planned to have 3 photolithography equipment and glue development equipment. The total number of equipment is less than 100. It is actually a miniature version of the Jiading production line. After the equipment is verified, it can be pulled to the large production line.
btw, that's 2 billion RMB for a domestic production line. This should be the tip of iceberg for how much they are investing for domestic tools around China. His comments are a little conflicting because he talks about already have a 28nm line with domestic line that an eventually be extended to 14nm, but then later makes it sounds like they haven't started testing it yet.
My guess is that what they have right now is a fully functional Arf dry that's been operating for a few months (remember the LAM employee interview from Oct where he said ICRD Jiading has a fully Chinese tools line) When we hear recently about the few domestic tools that just got verified for Arf process, that most likely got verified in Q4. We are almost reaching the point where they are confident about it, so mass production for Arf should start in a couple of months, I guess.
Again, ramping up production of everything is key here.
星海: 全国产的28nm产线跑通了吗?如果没跑通,是不是只差光刻机了?
havok: 没有。还要2,3年吧
Xinghai: Has the domestic 28nm production line operational already? If it doesn't , is lithography machine the bottleneck ?
havok: Not yet. Another 2 to 3 years needed
Keep in mind that this was in reference to ICRD. When asked about why it takes so long, he answered.
别的设备材料先不论,仅光刻机要让他在产线稳定量产一般就需要几年,当然因为上微90nm干式机已量产,浸没机与干式机在结构上无太大改动所以验证时间会相对快一些
In general, I feel like he is very reluctant to talk about progress outside of SMEE and ICRD, since he is mostly familiar with the work they are doing with ICRD. Again, that's why I think DUVi will get validated on SMIC first.
His other points are quite enlightening:
1) It implies that there are non-lithography scanner tools that are still need to be fully developed and validated. We may have seen patents and such announced, but that does not mean certain tools are ready for a production line. I think we can assume that they can get these tools for other countries like Japan, Korea and Netherlands. Whether they can be utilized for 14nm production is another story, but I think 22 to 28 nm should be no problem.
2) The first lithography machine is always the hardest. It took them 4 years from that original certification of SSA600 before it started mass production. So his comment about a few years is likely based on that painful experience. It's reasonable to assume that SSA800 dry has taken a lot less time to reach this point, since it's almost ready for mass production.
3) He is expecting another 2 to 3 years for ICRD to go fully domestic on DUVi (keep in mind that no other Chinese fabs need to go fully domestic, so that's not a consideration for them), because it will take them a few more months to fully validate the 55nm all domestic production line. Which means, it will take another 2 years on top of that to move in the domestic tools for 28nm production and then validate it. That seems a little longer than expected, but is not crazy.
I think SMIC is likely a lot better at this stuff than ICRD and will get their validation of SSA800 immersion finished sooner than that. I think early 2024 (so a little over 1 year in validation of process) is entirely reasonable. I would be surprised if SMIC doesn't have a couple of SSA600 in its production line somewhere, so they should have some experience already with SMEE front end scanners.
My guess is that the second gen DUVi is also likely to get validated in SMIC first with a process that's de-Americanized, but not fully domestic. Again as of now, I predict early 2024 to start mass production of SSA800 immersion and early 2025 for second gen DUVi that will start of being 14nm and eventually used for 7nm.
SMIC needs to import enough of the later generation ASML DUVi scanners to last until late 2024 if they do not want to miss any additional time in their advanced process production.
I remember CETC working on DUVi too. Don't know the progress.
But the underlying suppliers are the same, so unless they can get that ramped up, there is an upper limit to the number of DUVi they can produce.