At this year's International Conference on Extreme Ultraviolet Lithography, in Monterey, California, a Plenary paper was featured by Harry Levinson and Timothy Brunner, titled "Current challenges and opportunities for EUV lithography." It is available as an open access paper: Current challenges and opportunities for EUV lithography
It nicely lists the main issues that are being addressed today:
1. Lack of pellicles
2. Stochastics
3. 3D mask OPC complications
4. New resists needed
5. Aberrations
6. Scattering (Flare, Out-of-band)
7. Degrading source efficiency
8. Uptime/downtime
In particular, for the lack of pellicles, there is no longer an allowance for dark-field via layers. Instead, it is mentioned that contamination can frequently occur during transport and usage, to the point that procedures are necessary for keeping good yield.
A flow was provided where specially prepared wafers (separate from product) which use the same reticle but are designed for better defect sensitivity during inspection, are exposed and then inspected for repeating defects.
If there are no repeating defects, the reticle is considered clean.
Otherwise, the mask is cleaned, and the exposures and wafer inspections are repeated. The mask cleaning would also be repeated as necessary. In addition, if any product wafers were already exposed by the suspect reticle beforehand, they need to be inspected and possibly reworked.
Thus, a dirty reticle would need to be exposed in at least three rounds, with the third or final round for product (the first exposure discovers a dirty reticle from repeating defects, and at least a second exposure is needed to confirm it is clean with no repeating defects). This is obviously a significant productivity hit.
The impact is larger with increasing inspection frequency.
So pellicles continue to be a high priority issue. Also, inspection of pellicles, which also needs to be developed, would help in this area.
It nicely lists the main issues that are being addressed today:
1. Lack of pellicles
2. Stochastics
3. 3D mask OPC complications
4. New resists needed
5. Aberrations
6. Scattering (Flare, Out-of-band)
7. Degrading source efficiency
8. Uptime/downtime
In particular, for the lack of pellicles, there is no longer an allowance for dark-field via layers. Instead, it is mentioned that contamination can frequently occur during transport and usage, to the point that procedures are necessary for keeping good yield.
A flow was provided where specially prepared wafers (separate from product) which use the same reticle but are designed for better defect sensitivity during inspection, are exposed and then inspected for repeating defects.
If there are no repeating defects, the reticle is considered clean.
Otherwise, the mask is cleaned, and the exposures and wafer inspections are repeated. The mask cleaning would also be repeated as necessary. In addition, if any product wafers were already exposed by the suspect reticle beforehand, they need to be inspected and possibly reworked.
Thus, a dirty reticle would need to be exposed in at least three rounds, with the third or final round for product (the first exposure discovers a dirty reticle from repeating defects, and at least a second exposure is needed to confirm it is clean with no repeating defects). This is obviously a significant productivity hit.
The impact is larger with increasing inspection frequency.
So pellicles continue to be a high priority issue. Also, inspection of pellicles, which also needs to be developed, would help in this area.
Last edited: