Intel FAB 52/62 are High-Na capable

[siliconbruh999]

https://twitter.com/x/status/1980502598085148955

 

[Xebec]

According to ASML on X:

Time to nerd out! Here are some key specs of our High NA EUV system:

Imaging resolution: 8 nanometers
Productivity: ≥185 wafers per hour (at 20mJ/cm²) with a roadmap to >220 in 2025
Optics type: Anamorphic mirrors
System dimensions: 14 m x 4 m x 4 m
System weight: 150,000 kg

(Just over 13 feet tall for those of us raised in a number system that makes no sense).

 

[Daniel Nenni]

According to ASML on X:

Time to nerd out! Here are some key specs of our High NA EUV system:

Imaging resolution: 8 nanometers
Productivity: ≥185 wafers per hour (at 20mJ/cm²) with a roadmap to >220 in 2025
Optics type: Anamorphic mirrors
System dimensions: 14 m x 4 m x 4 m
System weight: 150,000 kg

(Just over 13 feet tall for those of us raised in a number system that makes no sense).

How many HNA-EUV systems will they need per fab is my question? If we know the layout of the fab we could probably do the f=math here.

 

[bilau]

According to ASML on X:

Time to nerd out! Here are some key specs of our High NA EUV system:

Imaging resolution: 8 nanometers
Productivity: ≥185 wafers per hour (at 20mJ/cm²) with a roadmap to >220 in 2025

so is it:

185 x 24 (hr) x 30 (days) = 133K patternings per machine-month, (max.)?

 

[Xebec]

so is it:

185 x 24 (hr) x 30 (days) = 133K patternings per machine-month, (max.)?

Not an expert, but that is my understanding. The theroetical maximum throughput (0% downtime, single exposure/pattern) would be that, using 2024 specs for the system. For comparison:

 

[Fred Chen]

According to ASML on X:

Time to nerd out! Here are some key specs of our High NA EUV system:

Imaging resolution: 8 nanometers
Productivity: ≥185 wafers per hour (at 20mJ/cm²) with a roadmap to >220 in 2025
Optics type: Anamorphic mirrors
System dimensions: 14 m x 4 m x 4 m
System weight: 150,000 kg

(Just over 13 feet tall for those of us raised in a number system that makes no sense).

20 mJ/cm2 is too low. That's for the development-only EXE:5000, which is actually discontinued (5 systems total). The EXE:5200 dose was upped to 50 mJ/cm2 (175 WPH).

 

[Fred Chen]

How many HNA-EUV systems will they need per fab is my question? If we know the layout of the fab we could probably do the f=math here.

18A pitches are too large for High-NA. Or, another way of putting it, if they say they are using single exposure with Low-NA, then the High-NA is disadvantageous due to one-third depth of focus. If they're planning 14A in Fab 52 asap, it sounds like 18A is relatively short-lived.

 

[siliconbruh999]

18A pitches are too large for High-NA. Or, another way of putting it, if they say they are using single exposure with Low-NA, then the High-NA is disadvantageous due to one-third depth of focus. If they're planning 14A in Fab 52 asap, it sounds like 18A is relatively short-lived.

I don't think 18A is short-lived otherwise their CEO/CFO won't be screaming 18A is a long term node.

14A is going to happen now I think.

 

[Fred Chen]

14A is going to happen now I think.

It would be happening in Oregon first, but how much has really happened.

 

[siliconbruh999]

It would be happening in Oregon first, but how much has really happened.

I am pretty sure LBT's Chess move worked and as for how much well only Intel knows.

 

[Fred Chen]

Resist blur is more important than the EUV NA.