Chinese semiconductor industry

tokenanalyst said:

UV-LED may replace UV Lamps in some lithography applications but high power DUV excimers lasers are necesary for high performance frontend lithography patterning.

Click to expand...

gelgoog said:

Like some here have talked about, performance of UV-LED light sources has improved enough in recent years that it might make sense to use that as the light source for older nodes. That would be a major disruption to the market if they got it to work. Since UV-LED is a lot more efficient at generating light than an excimer laser or a mercury arc lamp.

Click to expand...

problem with UV-LED is that it is very hard to generate short wavelengths because photon emission energy is the energy of the direct bandgap.

The larger the bandgap the higher energy the light emitted, but the larger the bandgap, the more insulating the material is because the room temperature concentration of charge carriers declines exponentially with increasing bandgap.

The worst part is that with bandgap decreases with temperature for semiconductors, so poor efficiency = heat generation = changes the color, making it not monochromatic = poor optical performance.



KrF excimer laser is still lower wavelength (248 nm). The wavelength is temperature independent since excimer dissociation energy is constant, and its monochromatic.

@horse Bro after reading your post and backtracking to 2021, most of our prediction had come to fruition, it is a year late BUT boy it was fun. It click All the boxes.

The improvement on SSA800 DUVL (same level with NXT2000i)- CHECK.

7nm chip program (already install in Huawei Mate P60 Pro) - CHECK.

Enhance Power source for DUVi (100w confirm) and EUVL (a LPP in 2024 and SSMB in 2025) -CHECK.

ansy1968​

Brigadier​

Registered Member

• Dec 20, 2021
• Add bookmark
• #9,963

Since we are ending the year, we can say China IC had achieved a lot, 2020 will be remembered as the start of Chinese IC renaissance. In 2021 almost all spectrum of manufacturing indigenous 28nm and 14nm chips will be achieved, not leading edge BUT at least self sufficient. While in Chip Design we are on par or leading in some areas as stated below, (will not post the YT video, members that is interested please see the inscription below) It also include packaging and 3D stacking tech.

Some areas needed to focus on in 2022

1)Improvement on SSA800 DUVL to ASML NXT 2000i -2050i level.

2) 7nm chip indigenous program.

3) power source for both enhance DUVL and EUVL program

ansy1968 said:

@horse Bro after reading your post and backtracking to 2021, most of our prediction had come to fruition, it is a year late BUT boy it was fun. It click All the boxes.

Click to expand...

Yeah, brother ansy1968, those were interesting days.

The foreigners talking the same points now in their articles. For the EUV it is the workbench, the light source, the lens. Those were the hurdles, among many of them.

Had this other idea a day ago.

They are actively into building an ASML type of EUV, and they are actively looking into the SSMB method which is a big project.

Why do both?

This is the part we have to remember who is doing this, it is the government, and the Chinese government is cautious. The point here, is the government is not a business, therefore they will take no chances. Does not matter how much money it takes. Risk is not a factor here.

So, logically, if they want to get this EUV done, and take no chances, then one project is the primary one, and the other project is the back up.

Which is which, seems kind of obvious, depending on who we talk, heh!

This is going to be very interesting.

My impression is, speaking as someone who does not know much about this stuff, that Kirin-9000s is the world's best 7nm chip and can give a 5nm a run for its money.

If you were a manufacturer, why bother using the 5nm chip when the 7nm is almost the same, and cost less?

But there, that brings up an interesting point. Like how comrade samwenyu posted in another here, the purpose of the EUV in China, is not to get to 5nm, it is going for 3nm and 2nm. Why not. That's business. You want to beat the competition to the punch.

So this story, of EUV in China, seems like one of those epic Chinese stories or moments, where we probably just a quarter or a third of the way into it.

Then the tech, will intersect with the business and economics, and the politics. Like everything. Like a full cage match!

Wuhoo!

BlackWindMnt said:

We were talking about ssmb in 2021 feels good to be early even if it was just hopium back then.

Click to expand...

To be frank it still is hopium. Two years later, and there's still almost nothing. Just some experimental papers and an experimental facility under construction.

If you really believe this magical SSMB EUV will be mass-producing 5nm chips in 2024 or 2025 - I have some Tejas stealth jet fighters to sell you.

By the way, according to this thread 28nm DUV was supposed to be here this year. Where is it?

Interesting read This lady is not as optimistic as Gina R. She see the tide receding and the tsunami of Chinese chips...

ansy1968 said:

@horse Bro after reading your post and backtracking to 2021, most of our prediction had come to fruition, it is a year late BUT boy it was fun. It click All the boxes.

Click to expand...

Oh yeah, one other thing brother ansy1968, that we should definitely mention.

This thread predicted that they will announce the DUVi and EUV in China together at the same time, or roughly the same time.

Well, look at what happened.

US Commerce minister comes to China, the person mainly responsible for all these chip bans.

The Mate 60 Pro suddenly appears.

Now before it official debut date, some article on the Chinese internet about EUV, actually made the rounds in English alternative media, and made it into the South China Morning Post, which is important, because there is where all the American spies are, in Hong Kong. According to legend, of course, I don't want to get arrested myself.

What the funny thing is, is that Huawei still would deny everything, and will not even claim the Mate 60 Pro is a 5G phone.

It is a non-announcement announcement. We know it's coming.

Hadoren said:

To be frank it still is hopium. Two years later, and there's still almost nothing. Just some experimental papers and an experimental facility under construction.

If you really believe this magical SSMB EUV will be mass-producing 5nm chips in 2024 or 2025 - I have some Tejas stealth jet fighters to sell you.

Click to expand...

Well, back in 2021 SSMB was a bit more theoretical than it is now. Since then, further reinforcement towards progression to building the trial facility and indicators that they are committing to both SSMB as well as LPP, is what I would call progress.

However, if anyone in 2021 or in 2023 believes that SSMB EUV would be ready for producing chips in 2024 or 2025, then obviously that's unrealistic. If SSMB EUV does have practical utility for producing chips, then it likely won't be seeing anything resembling commercial use until late this decade (at the earliest)


But there definitely has been progress with SSMB, from the perspective of how something that is academic and research in nature can transition to something that is more experimental and developmental in nature, which is how I have observed the progress.
If one wants a "MOAR CHIPS NOW" or "WHERE EUV" without actually tracking the details of it, then sure you're going to be disappointed.

Hadoren said:

By the way, according to this thread 28nm DUV was supposed to be here this year. Where is it?

Click to expand...

As far as we know the first products should be in the hands of the customers.
Chances are it won't be until a couple of years later until we know in retrospect that "oh, SSA800 had actually begun trial production on XYZ date".

horse said:

My impression is, speaking as someone who does not know much about this stuff, that Kirin-9000s is the world's best 7nm chip and can give a 5nm a run for its money.

If you were a manufacturer, why bother using the 5nm chip when the 7nm is almost the same, and cost less?

But there, that brings up an interesting point. Like how comrade samwenyu posted in another here, the purpose of the EUV in China, is not to get to 5nm, it is going for 3nm and 2nm. Why not. That's business. You want to beat the competition to the punch.

So this story, of EUV in China, seems like one of those epic Chinese stories or moments, where we probably just a quarter or a third of the way into it.

Then the tech, will intersect with the business and economics, and the politics. Like everything. Like a full cage match!

Wuhoo!

Click to expand...

Bro Liang Mong Song had laid out the Chip development for SMIC, remember his famous resignation letter. Both the 5nm and 3nm nodes had finished their development and are waiting for the necessary equipment. That was in 2020, fast forward now, I think SMIC N+3 5nm is on its way (using NXT 2050i), we may see it on Huawei Mate 70 pro next year. and the 3nm nodes by 2026, like I said before only the Chinese (Beijing and Taipei) will be producing both 3nm and 2nm chips....lol

Liang Mong Song: SMIC's 7nm node has already completed full development, and is preparing for mass production in April 2021. Designs and most of the development for 5nm and 3nm nodes have also already been completed, only awaiting EUVL delivery at this point.​

Hadoren said:

To be frank it still is hopium. Two years later, and there's still almost nothing. Just some experimental papers and an experimental facility under construction.

If you really believe this magical SSMB EUV will be mass-producing 5nm chips in 2024 or 2025 - I have some Tejas stealth jet fighters to sell you.

By the way, according to this thread 28nm DUV was supposed to be here this year. Where is it?

Click to expand...

We will be enjoying ssmb EUV blasted silicon chips before India will be flying a fully domestic Tejas stealth jet.

I do think China will have some sort of EUV capability either in trial or low volume production before the start of 2026. Because EUV capabilities will be the cherry on top of the already successfully baked "made in China 2025" cake. I have been saying this since I joined this forum.

/hopium

BlackWindMnt said:

We will be enjoying ssmb EUV blasted silicon chips before India will be flying a fully domestic Tejas stealth jet.

I do think China will have some sort of EUV capability either in trial or low volume production before the start of 2026. Because EUV capabilities will be the cherry on top of the already successfully baked "made in China 2025" cake. I have been saying this since I joined this forum.

/hopium

Click to expand...

However the first EUV capability that they get will almost certainly not be in the form of SSMB, which is the claim in question, and it is also why a few pages ago I said that it's stupid that news media are reporting about SSMB as if it is China's primary strategy in pursuing EUV lithography.

The actual success or timescale of EUV efforts will have no relation to political significance or relating to publicly announced policies like MIC 2025.


There's really no need to view things in any manner apart from an analytical lens.