Making holes in the Si wafer via DRIE or holes in SU8 and using it as negative mold, and cure PDMS on it to get PDMS pillars?
or
Making pillars from SU8 or Si (DRIE) and making a negative mold from PDMS from this. Then using PDMS negative mold to get PDMS pillars?
Thanks
This greatly depends on the aspect ratio and size of pillars you want. We have produced pillars with aspect ratios as high as 20:1 with direct casting and double casting with SU-8 with great success as long as the post diameter is >2 microns. We used RIE Si to make 1 micron or smaller posts but it is very expensive and the side profile of our posts was variable which makes it harder to characterize. Also the PDMS would break off inside the holes in the Si and it is virtually impossible to dissolve it out which means that you have to replace your masters frequently (expensive).
Thank you very much for the information Ron. For SU8 direct casting, I was not very successful. What are your steps to do it?
for SU8 double casting, when you make your su8 pillars, did you use a primary layer of su8 to increase sticking of the pillars to wafer?
SU-8 is a tricky resist so I'll tell you a couple of my tricks to improve the performance of my casting masters in general. In my experience, It is important to do a much longer soft bake than the datasheet suggests. Unlike most novolak resists, there is no chemical process happening during this bake, you are just baking out the solvent which increases the Tg. This bake can be an hour long and it won't hurt a thing. On the other hand, if you don't bake long enough, the solvent will act as a photo-inhibitor and since it is hard to control the solvent content exactly, you will have exposure variability. I find that when the resist is thoroughly baked out, the best exposure dose is 10% below the bottom exposure line on the datasheet. If the exposure dose is too high, you will have adhesion failure.
Another thing that is suggested by the datasheet is that you use two hot plates, one at 65oC and the other at 95oC. The idea behind this is that is will help reduce film stress. I say HOGWASH! What they neglect is that you will get most of your film stress on the cool down as you are crossing the Tg. They make no suggestions about what to do when you take your substrate off of the 95oC hotplate.
My method uses a hotplate with a ramp function. On the soft bake you can put the substrate directly on the hot plate. Your resist will spend at least 5 minutes at 30oC above the Tg which is plenty of time for the stress to relax out. Then allow the substrate to cool to 40oC on the hot plate. This takes a longtime but will greatly improve adhesion. Then on the post exposure bake, set the hot plate to ramp at 3oC per minute once you reach 95oC you will dwell for the time suggested on the data sheet for your resist thickness and then turn off the hotplate. Allow to cool as in the soft bake. Develop as usual. Hard bake should not exceed 150oC and the ramp should be no greater than 2oC per minute.
Next, you will want to silanize. It is important not to over treat the SU-8 in plasma since it will begin to ash and the silane will not stick to the bulk. 20 seconds in a Harrick plasma cleaner is plenty (higher power plasma ovens may take less time). Then use a fluorocarbon silane to treat the surface.
This procedure should produce casting masters that are both durable and easy to part from your elastomer.
Good Luck!
I agree with Ron, SU8 takes time and you have to take care with cool down steps.
Something I don't undestand is your 2nd option, using PDMS mold to produce inverse PDMS structure. it seems to me that curing PDMS on cured PDMS will produce a compact bloc of PDMS.
I would suggest using a negative SU-8 pattern, silanize before PDMS casting and gently extract the cured PDMS from the mold
PDMS on PDMS casting needs a silanization step for the PDMS casting master. We use vapor deposited Perfluorooctyltrichlorosilane as a release layer.
Using an Si master should last far longer than a comparable SU8 mould, though you need to take care to passivate the surface in both cases for repeatable demoulding. I strongly prefer using DRIE on Si, as I can arbitrarily set my height within half a micron or so - in contrast to SU8, whose height may vary strongly across the width of the wafer - and finishing the process with a a deposition step (with SF6) removes the need to perform a separate silanization step, as with SU8. Be wary, though, that for very small post sizes (10:1).
Indentation lithography is the best technique to fabricate nanopillars. For more information, please find attached.
- Badges
- Science topic