ASML H2 2021 EUV update at SPIE

[Fred Chen]

At the SPIE EUV conference a few months ago, ASML's Matthew Graham presented a brief update on the status of EUV tools.

Briefly:
1. In Q2 2021, 92-102 tools exposed ~11 million wafers (although peak output 2500 WPD)
2. Availabiliy ranged from 75% to 95% by end of Q2 2021
3. NXE:3600D increased throughput >18% over NXE:3400C, giving 160 WPH @30mJ/cm2
4. EUV collector lifetime improved to 10% drop after 200 billion 50kHz pulses (~46 days).
5. In the research lab, up to 500W achieved, but customer use is at 250W.

Paper (presentation only): D. C. Brandt, I. V. Fomenkov, and M. Graham, "Performance and availability of EUV Sources in high volume manufacturing on multiple nodes in the field and advances in source power," Proc. SPIE 11854, 1178540J (2021).

 

[Tanj]

Does that mean the collector needs refurbishing or swapping every 46 days (100,000 wafer exposures)? If 20 exposures are needed for logic, this cost overhead averages 1 per 5,000 wafers? Seems likely a significant cost.

 

[Fred Chen]

Does that mean the collector needs refurbishing or swapping every 46 days (100,000 wafer exposures)? If 20 exposures are needed for logic, this cost overhead averages 1 per 5,000 wafers? Seems likely a significant cost.

They can compensate to some degree by slowing the stage down (to keep the dose the same), until the throughput is not acceptable.

 

[VCT]

At the SPIE EUV conference a few months ago, ASML's Matthew Graham presented a brief update on the status of EUV tools.

Briefly:
1. In Q2 2021, 92-102 tools exposed ~11 million wafers (although peak output 2500 WPD)
2. Availabiliy ranged from 75% to 95% by end of Q2 2021
3. NXE:3600D increased throughput >18% over NXE:3400C, giving 160 WPH @30mJ/cm2
4. EUV collector lifetime improved to 10% drop after 200 billion 50kHz pulses (~46 days).
5. In the research lab, up to 500W achieved, but customer use is at 250W.

Paper (presentation only): D. C. Brandt, I. V. Fomenkov, and M. Graham, "Performance and availability of EUV Sources in high volume manufacturing on multiple nodes in the field and advances in source power," Proc. SPIE 11854, 1178540J (2021).

Can EUV last 30 years like traditional Lithography Equipment?

 

[M. Y. Zuo]

160 WPH gives a far higher WPD figure even at 75 to 90% availability. What could explain the difference?

 

[Fred Chen]

160 WPH gives a far higher WPD figure even at 75 to 90% availability. What could explain the difference?

Dose is not always 30. Peak WPD was 2500, but not average. So some tools virtually not moving.