Yup. The numerical aperture is not the major bottleneck here. That part is relatively easy to design. It’s the power of the light source that’s the bottleneck. The speed at which high NA can be deployed is dependent on how powerful your light source is. Alternatively, you can also reduce dependence on higher power light sources by developing optics with lower level scattering loss, but the materials science on this is probably at a limit right now. Photons that energetic are extremely difficult to handle.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
- Not open for further replies.
The higher image contrast of HiNA projection system compare to the current EUVL will provide more photon not less. This is one selling point of he HiNA system in addition to better resolution capability.
If ASML could really successfully deliver the HiNA EUVL then it would be even more difficult for SSMB to compete. The gap for us to catch up actually widens.
This is a quite deep article on problems with moving to Hi-NA in EUV
And this just it, the “equivalent optics” of HiNA is even more difficult than the current EUVL hence my last comment.From what I've read, high NA is optics based improvement, not source based improvement. So it is completely independent of the EUV source, SSMB with equivalent optics would have equivalent improvement.
You got it backward. If you don’t believe me, you can check what ASML says. According to ASML, the source is a small improvement. The major effort is in the optics.Yup. The numerical aperture is not the major bottleneck here. That part is relatively easy to design. It’s the power of the light source that’s the bottleneck. The speed at which high NA can be deployed is dependent on how powerful your light source is. Alternatively, you can also reduce dependence on higher power light sources by developing optics with lower level scattering loss, but the materials science on this is probably at a limit right now. Photons that energetic are extremely difficult to handle.
I mean, have you seen the metrology vacuum chamber Zeiss had to build to make the HiNA mirror? And going with anamorphic 4x/8x reduction mirror is also a big change.
Mirror size is *a* factor. And there are *some* challenges to making a larger mirror. But while not optimal you don’t absolutely *need* to create a larger mirror to get to a higher NA. This isn’t refractive optics, so nothing says your mirror *must* be single piece. Hence larger single piece mirrors actually aren’t the fundamental constraint.You got it backward. If you don’t believe me, you can check what ASML says. According to ASML, the source is a small improvement. The major effort is in the optics.
I mean, have you seen the metrology vacuum chamber Zeiss had to build to make the HiNA mirror? And going with anamorphic 4x/8x reduction mirror is also a big change.
But regardless of how you make your big mirror, a larger mirror=more scattering loss, because you’re projecting the fixed quantity of photons generated from your light source across a much larger area. It makes no sense to increase your aperture if you end up with fewer photons per area at your point of illumination. That’s the much bigger issue. Remember, this is not a passive illumination system but an *active* illumination system. It means you need to account for the quantity of light you are trying to project.
Marketers don’t like explaining the actual physics of what they’re selling.
Last edited:
Yes, larger aperture means you’re capturing photons from a wider angle, but first you need more photons to capture, and if scattering loss is a big factor, you’re going to need to generate more photons to compensate for wider projection area to keep your illumination per area equal or better than at lower NA. This is not just a minor issue. From a physics and engineering standpoint this is the *fundamental* constraint. Why do you think they’re going anamorphic to maximize the amount of photons reflected off the reticle? Hint, it has something to do with photon quantity and photon loss being the fundamental constraint.I’m not arguing you’d need more source power. It’s just not a big deal compared to the lens.
with larger aperture, photons from a wider angle will be collected and projected into the wafer, this is why the image contrast of HiNA EUVL compared to current EUVL will be higher. Anamorphic reduction mirror will be used to cut the scanning incident angle in half to maximize the amount of photon reflected off the reticle.
They went with anamorphic, a more complicated optics design, to try to reduce the need for an even more powerful light source. In other words, from an engineering design standpoint it was preferable to pursue a more complicated optical design than to pursue more powerful light source. That should tell you quite a lot about which one of these factors is the bigger challenge. To put it another way, if you have a more powerful light source you don’t require as complicated an optics design.the change from 4x reduction to 4x/8x reduction is a big change, a bigger issue than the source. Also a big change for the fabs as this is a big change from old ways. The stages also have to be faster to compensate for the smaller field size (as a result of limitation of reticle size and the larger 8x reduction in the scan direction).
like I said, all these changes combined require bigger effort than the larger source power ‘issue’.
It’s important to remember that the primary advantage of high NA is not that a higher NA means capturing more light, but that it allows for a beam with a wider angle of incidence, and that wider angle of incidence is what allows you to increase scanning resolution.
Last edited:
I’m also talking about fundamental principles of system design here. You can increase optical complexity to compensate for less sufficient light source, which is what ASML has done, *but* if you don’t have a constraint on the light source why would you choose to increase optical complexity? Or alternatively, if you think your development schedule for improving your light source can be faster than the schedule needed to design a more complex optical system, why would you choose the latter? Imo it makes absolutely no sense to talk about the projection system as if it’s independent from the light source.I think I see our differences @latentlazy I’m looking at the entire HiNA projection system as a whole when stating this is more difficult than the source. You are comparing just the HiNA mirror itself to the source.
@latentlazy, I agree with everything you said. But why would you comment about the constraints of the route that was not chosen for build-up of HiNA EUVL (keep system same except go with HiNA mirror and much higher source)? I thought we are comparing the HiNA EUVL design vs current EUVL.Yes, larger aperture means you’re capturing photons from a wider angle, but first you need more photons to capture, and if scattering loss is a big factor, you’re going to need to generate more photons to compensate for wider projection area to keep your illumination per area equal or better than at lower NA. This is not just a minor issue. From a physics and engineering standpoint this is the *fundamental* constraint. Why do you think they’re going anamorphic to maximize the amount of photons reflected off the reticle? Hint, it has something to do with photon quantity and photon loss being the fundamental constraint.
They went with anamorphic, a more complicated optics design, to try to reduce the need for an even more powerful light source. In other words, from an engineering design standpoint it was preferable to pursue a more complicated optical design than to pursue more powerful light source. That should tell you quite a lot about which one of these factors is the bigger challenge. To put it another way, if you have a more powerful light source you don’t require as complicated an optics design.
It’s important to remember that the primary advantage of high NA is not that a higher NA means capturing more light, but that it allows for a beam with a wider angle of incidence, and that wider angle of incidence is what allows you to increase scanning resolution.
Without you now going to details explanation so people now understand reason why ASML chose to make big changes to the optical system, I’m sure most people would have missed the points from what you said originally.
Dude, you are too theoretical. No one is asking for reason of the big change to anamorphic mirrors. If that’s what we are talking about then everything you said would be appropriate.I’m also talking about fundamental principles of system design here. You can increase optical complexity to compensate for less sufficient light source, which is what ASML has done, *but* if you don’t have a constraint on the light source why would you choose to increase optical complexity? Or alternatively, if you think your development schedule for improving your light source can be faster than the schedule needed to design a more complex optical system, why would you choose the latter? Imo it makes absolutely no sense to talk about the projection system as if it’s independent from the light source.
what we were talking about, at least that’s my interpretation, is on the difficulty for ASML to build their HiNA system.
like I said, I recognize and understood all that you said. But we weren’t really talking from the exact same topic. I misinterpreted what you tried to say earlier.
- Status
- Not open for further replies.