Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
I don't know if you read Chinese or not，but that's not the meaning of his word
ID "秦声楚舞" asks：不是说接近交付吗？ Translation： Didn't you say near delivery？
ID "havok" replies : 官宣交付呗，judging from the context，he is saying "just wait for official announcement about delivery"
Remember，there are some posts in the middle got deleted by havok，don't know why he did that. For instance，this post was deleted by him，so you cannot find it now
ID "晓浩仔" ：what is the status of domestic DUVi now？Already testing now？
ID "havok" answer to ID "晓浩仔" ：Near delivery. Deliver to fab（for testing），not mass production（yet）
Let's put a few things out there.
90nm in production -> that refers to SSA600. This was already known from earlier this year. Not sure how commercially competitive it is, but it's clearly in use somewhere (Yandong and maybe somewhere else) for them to refer to it as such.
The new SSA800 is expected to actually be commercially competitive. The dry version is actively getting tested in production line at ICRD. It should be ready for mass production soon. However, it sounds like they are constrained to only making 5 of them a year based on Havok's comments about one of the components.
DUVi work passed China's review last year. It was delayed due to CIOMP being slow with exposure system.
btw, CIOMP being slow in 02 project is a concern for the EUV project since it is also working on integrating a prototype right now for EUV project IIRC. IIRC, Shanghai Institute of Optics and Fine Mechanics is the other major player here.
Anyhow, he began his piece with
I think he is being deliberate here in not spelling things out perfectly clear. You could interpret "接近交付" as about to be delivered in the near term or it has been delivered recently. I personally don't think it matters that much either way. The point is that it will be ready to start validation with customer in the near future. Again, I think he is being very intentionally unclear here, because he doesn't want to get in trouble.
He is talking a lot right now, because he is proud the project is almost finished. He wouldn't be talking this openly if we are still a year away from start of validation.
90nm in production -> that refers to SSA600. This was already known from earlier this year. Not sure how commercially competitive it is, but it's clearly in use somewhere (Yandong and maybe somewhere else) for them to refer to it as such.
The new SSA800 is expected to actually be commercially competitive. The dry version is actively getting tested in production line at ICRD. It should be ready for mass production soon. However, it sounds like they are constrained to only making 5 of them a year based on Havok's comments about one of the components.
DUVi work passed China's review last year. It was delayed due to CIOMP being slow with exposure system.
btw, CIOMP being slow in 02 project is a concern for the EUV project since it is also working on integrating a prototype right now for EUV project IIRC. IIRC, Shanghai Institute of Optics and Fine Mechanics is the other major player here.
Anyhow, he began his piece with
which indicates that it shouldn't be taken as fact. I see his comment as indicating that it is somewhat like this and we shouldn't follow it word by word. If anything, I think he wrote this in earlier part of 2022, which means SMEE thought it was going to deliver a working prototype to customer for validation in 2022. Keep in mind that while we use the word prototype here, this is actually a machine that has already passed review. So we should think of this as a beta version of the production. Calling it just prototype makes it sound a lot less further away than intended to.
I think he is being deliberate here in not spelling things out perfectly clear. You could interpret "接近交付" as about to be delivered in the near term or it has been delivered recently. I personally don't think it matters that much either way. The point is that it will be ready to start validation with customer in the near future. Again, I think he is being very intentionally unclear here, because he doesn't want to get in trouble.
He is talking a lot right now, because he is proud the project is almost finished. He wouldn't be talking this openly if we are still a year away from start of validation.
He blames the huge delays on Changchun Institute of Optics (Guoke Precision)...nothing new here, business as usual
He didn’t blame all the huge delays on Changchun Institute and said they completed their task in 2020. You missed the other part about the delays related to the suppliers. It seems like the main issue isn’t the quality of the machine but issues with one supplier resulting in delays post 2020.
Another key piece of information:
Since the 28nm immersion lithography machine can be used to manufacture 14nm process chips, the conditions of multiple exposures can also forcefully manufacture 7nm process chips, but the cost-effectiveness is not high.
May be a bit optimistic as other posts seem to suggest that the first model of the ARFi won't be accurate enough for sub 28nm nodes.
Think the English the could use some clarification here.
“The 28nm immersion front-end lithography machine is the last critical task of the 02 special project. In June 2021, the overall technical plan and initial sample of the 28nm immersion front-end lithography machine of Shanghai Microelectronics has passed the expert review of the 02 special project.”
“Initial sample” is a translation of 初样, but it‘s probably referring to an initial *prototype*, or what is sometimes called a pilot instrument.
“According to the latest news from the company, all the previous technical problems have been solved. Although the epidemic has delayed some time, the company is very confident in the delivery of the prototype of the 28nm immersion front-end lithography machine within this year.”
The prototype they’re talking about delivering in 2023 is probably more specifically a production prototype. 交付 is the term used for “delivery” here and specifically refers to handing something off, the implications of which is that these prototypes aren’t for internal development, but are meant for customer use (probably for remaining verification, validation, or process integration development).
Havok’s post makes no mention of CIOMP being slow. The delay was due to the instrument base supplier encountering some technical difficulties.
I think this is pretty clear
The first part says the exposure system is very complicated, required the highest level of optical studies and took until 2020 for CIOMP to complete, had to keep testing and adjusting.
Originally intend to deliver prototype in 2021, but due to supply issue with base of lithography machine, missed the 2021 delivery window. Seems to me this is not related to CIOMP. And also seems to me that the 2021 delivery was referring to production prototype to customer. Now they have either just delivered it to customers for production validation or about to.
According to the latest news from the company, all the previous technical problems have been solved. Although the epidemic has delayed some time, the company is very confident in the delivery of the prototype of the 28nm immersion front-end lithography machine within this year.
as @tphuang mentioned, once you achieve success in 28nm DUVi. advance variants will come very quick.
so SMEE finally done it.
This has gone exceedingly well.
To me, it is the timeline, and the nature of the IC industry.
The nature of the IC industry, is that everything must move in lock step. All components that are needed, must be about the same standard. Such as the photoresist is a different chemical for more lower nodes, and it is more complicated for higher nodes.
The entire industry, must move all together forward one step at a time.
That is why all the news posted in this thread is really important. If one component of the IC supply has advanced to a certain level, we can bet that other parts will be there too. Otherwise, there would be no progress. Everything must advance together.
The timelime is ultra important. It probably takes 3 to 6 months to build one of these machines.
But, they never built one, without foreign parts, so things had to change in sourcing, and perhaps in the factory itself that makes the machine. To do the sourcing for domestic components that work and fit, then to rejig the factory if needed, that takes time too. Maybe half a year, or more. We do not know.
So it might take one whole year, to make this SMEE lithography machine, when all the other preparatory work had been done.
The preparatory work, like sourcing domestic parts, designing the SMEE machine, that takes time too.
So it seems to me that the preparatory work was completed sometime last year, and it will take a whole year to build this machine by 28nm immersion water machine by SMEE.
Damn, that really is impressive.
Look at the timeline of Huawei sanctions, which were 2019. So these guys at SMEE had to start working on this project on 2020 will some urgency. But the pandemic hit, which probably is another 6 month delay.
2019 Huawei sanctions. Then need time to do all preparatory work for domestic components and blueprints design, probably a year. Then to make the machine, first time trying probably a year. Add in the pandemic, 2020, 2021, blueprints, assembly, 2022 delays from pandemic, man, this is freaking fast!
- Status