thx for the info Joe. I'll try to bake only the PDMS mold to rule out the
temperature factor. However, if I understand ur theory correctly, with PR on the
top side, PDMS mold should bend like a up-side-down bowl. but what I got is the
other way around, i.e. the side without PR is actually the side which expanded.
any way, thx for the info, I'll let u know if I get this tackled.

shao

--- On Mon, 3/30/09, jgrogan@seas.upenn.edu  wrote:


From: jgrogan@seas.upenn.edu 
Subject: Re: [mems-talk] PDMS mold bends after 100C baking
To: "General MEMS discussion" 
Date: Monday, March 30, 2009, 1:19 AM


-----Inline Attachment Follows-----


I don't think it's bending from the temperature. It's bending from the solvent
in the SU8. The same thing happened to me when I tried to spin coat 1813 on
PDMS. PDMS swells in solvents. If you drop a block of PDMS into acetone for a
couple hours the whole thing will swell up. Harsher solvents like toluene will
make it balloon up to 2x its original size. People take advantage of this to
remove unbound monomers in the PDMS so that hydrophilic surface treatments last
longer. The bad news for you is that if you spin coat photoresist on one
surface, that surface will swell up while the other surfaces remain the same,
and the result is a bowed substrate. Unfortunately, I don't have a good solution
for you because I didn't pursue it any further. Two possible suggestions,
though, would be to either:

a) coat the surface of the PDMS with something to prevent the photoresist
solvent from contacting the PDMS material. Maybe sputter or evaporate a metal
layer.

b) oxygen plasma treat the heck out of the surface. It's my understanding that
extended oxygen plasma creates a glassy layer on the surface of the PDMS, which
is why you shouldn't over do the plasma treatment when bonding pdms. But maybe
you could do it on purpose and form a surface layer to protect the rest of the
material.

good luck!
Joe Grogan