I have the wafers (small pieces ~1/4 of 2") with simple topography of 10 um in height. To make airbridge connections between some pads with spans about 15 um I need to make a kind of planarization steps. After that I plan to spincoat a kind of liftoff layers system (Copolymer/PMMA) to expose it via Ebeam tool and make metallization.
I have PMGI SF9, LOR10B and some traditional CAR-based PRs to form "thick" layers.
What kind of approach or strategy can you suggest for "flooding" all the valleys over the topography? I also can open the windows in the resist on the pad tops to obtain contacts, but this layer should be as thin as possible.
Thank you!
Please clarify on the material of the topography which is 10um in height.
Is it multilayered PR or some thing else?
You want to reduce it from 10um to 1-2um which means you want to etch it with good control over the height and then do e-beam litho for metallization.
But unless you tell which is the material of the topography it is hard to conclude which method to use to etch it.
Unless, I got the problem statement incorrect .
Dear Yaksh Rawal,
The topography is a deep etched GaAs based structure with Au based contact layers on top. I need to make metallic airbridges between some of the pads. As general approach it is needed to fill the "valleys" with sacrificial/planarazing material (resist) to form an airbridge temporary support. At some final step I will open the valleys again (remove the material) having the metal bridges suspended between the pads. So I cannot modify initial topography in irreversible way.
The pad-to-pad distance is about 10-20 um only.
P.S. It is not mandatory to make full planarization. 1-2 ums in height difference will be good enough by my opinion.
Hi Rinat,
"I have PMGI SF9, LOR10B and some traditional CAR-based PRs to form "thick" layers."
You can use single or multilayer of these PR. you need to optimize the recipe for spin speed and exposure dose, which you can do on dummy samples and once the recipe is fixed and checked for repeatability, use it on actual sample to fill in the valleys.
Hope it helps. :)
The question is in what direction I should go :)
As for now I see strong dependance of the top layer thickness on the actual position on the sample relative to distance to the rotation center during spin process and relative to local pads layout, e.g. three adjacent pads of different areas could have resist thicknesses more than 4-fold compared to each other. It is significantely complicates the top areas removal or levelling while having good fill of the valleys...
Maybe the question should be "how to spincoat thick resist layer using rather thin resist solutions with uniform distribution of thicknesses on tops of the topography".
While spinning, the center will have more thickness compared to edges.
But as the sample size is 1/4th of 2inch Wafer, the delta in the thicknesses will not be much. this you can measure using ellipsometry measurements on a dummy samples.
Lower spins speeds will provide thicker PR coatings, then you can post bake it to harden if required.
In general it is so, but I see strong but local deviations over the topography features. It is probably to make it more conformal, maybe a sequence of short processes at high speed?..
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Semiconductor