Hi, All:

   So what kinda softbaking process u guys using for ultra-thick SU-8 layer. I'm
working with 1mm thick SU-8. My process is:

Spincoating first layer(around 500um), softbake at 65C for about 2 hours and
slowly ramp to 110c, hold for 6 hours and slowly ramp to room temp. for the
first layer I can bake out aroun 28% of the solvent which SU-8 50 should have
31%. But the remaining 3% seems not a big problem and I can finish my first
layer exposure pretty well.

for the second layer, I use almost the same process, but hold at 110c for 8hr,
but I can only bake out around 24% of solvent. Because of that, during my PEB
process, severe wrinkle happen around 50-60C, although it will smooth out after
a while, some of my fine structures r destorted by these wrinkles. I tried to
lower the baking temperature to 90C and bake for 15hrs, but still, only 24%-25%
of solvent can be bake out.

Any one has a cure for this prolem? thx a lot

Guocheng Shao

--- On Mon, 12/1/08, Oakes Garrett  wrote:

From: Oakes Garrett 
Subject: Re: [mems-talk] very thick layer of SU-8
To: "General MEMS discussion" 
Date: Monday, December 1, 2008, 12:48 PM

Hi Aimi,

First some background, we have coated, exposed and developed 500 µm films of
SU-8 on 200 mm wafers.  We achieved the 500 µm coatings via a 2x 250 µm
coating process.

As others have already mentioned, solvent control is a critical issue.  You
will need to carefully extract the solvents.  There are risks or crusting the
top surface of the resist layer or even beginning a thermal cross-linking of the
resist if you try to bake the resist too quickly or for too long.  I liked the
idea that one member suggested of simply letting the solvent evaporate.  I would
add one thought: vacuum might be the way to go -- especially given a 2000 µm
film thickness.  Hopefully you have enough time for this step!!

Returning to the issue of thickness, I am not sure that an 8x 250 µm spin or
spray coating process is going to be the most successful method.  Again, one
member suggested that you could simply pour the resist onto the wafer and
perhaps use a LEVEL hotplate to spread the resist.  I think that this might be
your best approach.  Assume that surface tension will keep the resist on the
wafer and that (pi)(r^2)(h) is a good approximation of the resist volume.

Exposure with wavelengths greater that 350 nm is a must.  SU-8 is very
transparent to >350 nm wavelengths.  It is questionable whether these
wavelengths will penetrate 2 mm of resist.

Developing is going to be difficult as well.  If your features are large you
can dunk develop the wafers but this will probably take in excess of five to ten
hours!  Megasonic assistance is probably a must for this resist thickness.

I urge you to contact MCC for assistance with your project.

Lastly, best of luck with your project and let everyone at MEMS-Talk know if
you succeed.

Best Regards,
Garrett Oakes