Zygo Interferometry of Echelle Gratings

This deck illustrates the need to deproject interferometry of echelle gratings to understand the origin of instrumental artifacts in the device development process.

Transcript

1. zygo interferometry of gratings michael gully-santiago | grad student UT Austin Silicon Diffractive Optics Group
2. who michael gully-santiago astronomer, engineer, bicycle rider amanda turbyfill astrophysicist, basketball enthusiast
3. Interferometry tells us two things Where did we go wrong? Is it good enough?
4. Interferometry tells us two things Where did we go wrong? Is it good enough? Astronomers care about this Engineers care about this
5. There are two data products Where did we go wrong? Is it good enough? Astronomers care about this Engineers care about this PSF FULL MAP
6. PSF FULL MAP The two data products are related to each other
7. PSF FULL MAP The two data products are related to each other Sub aperture FFT € ∫∫
8. Let’s talk about the full map. Since we’re engineers
9. You have to de-project the full map.
10. de-project
11. We didn’t used to de-project
12. We didn’t used to probably because astronomers don’t care. de-project
13. We didn’t used to probably because astronomers don’t care. but now we’re learning something, because de-projected images look like what engineers see. shadow of grating holder Shadow of grating holder Photoresist thickening at edge Periodic structures resembling Newton’s fringes Comets from photoresist spin Photoresist thickening at edge de-project
14. We can do even better with shadow of grating holder Shadow of grating holder Photoresist thickening at edge Periodic structures resembling Newton’s fringes Comets from photoresist spin Photoresist thickening at edge
15. high pass* . We can do even better with shadow of grating holder Shadow of grating holder Photoresist thickening at edge Periodic structures resembling Newton’s fringes Comets from photoresist spin Horizontal artifact, probably from using the cart and slit. high pass filtering* is: way faster than fitting Zernike polynomials. k ~ 51 pixels is the median smoothing kernel d = d - median( d, k )
16. But since we are astronomers and engineers, we want the best of both worlds™, so: We are proud to introduce… ZIGZAG: Zygo Interferometry of Gratings with Zernike Analysis by Gully
20. H! L! !" Silicon! Desired line edge position! Realized line edge position! Value measured by Zygo!
21. I’m having a hard time getting the PSF to work. This is either hard or I am not smart enough to figure it out. Help!
22. In summary: gully@astro.as.utexas.edu | astronomer and engineer You have to de-project the full map. and high pass filter and ZIGZAG™ does it all except PSFs… for now.