Hi Qinghong,

AZ5214E can be a touchy resist in my experience; you do the same thing 2
days in a row, and get different results. Maybe we got a dodgy batch.
Anyway, the repeatability can be improved greatly by waiting about 5
minutes between every bake, expose and develop step. Also, the 5214
might be interacting with the other resist- try a longer bake for the
OCG825, and rest it for a while afterwards. I imagine you are flood
exposing the bottom resist- maybe bake it again and rest it before you
spin the 5214, if you aren't already. If you see any bubbling, or the
interface between the resists becomes cloudy, then this is probably the
problem.

If that doesn't improve your repeatability then maybe try one of these:
-A double layer of AZ2035, both layers spun at 3000 rpm, is fine for 2
microns of gold, but you need to get your conditions right to get a good
undercut.
-Put down a non-photodefinable polyimide that removes in PR developer
(e.g. PI2616, or else Brewer science sells a product called ProLIFT).
Then when you pattern your resist (any resist is fine), you
automatically pattern the polyimide with an undercut defined by the
'over-develop' time

-AZ4562 can be used for very thick liftoff: spin on a thick layer, bake,
flood expose, rest for ~1 hr, spin on a thin layer, bake, expose,
develop. The flood-exposed resist does the same job as the polyimide
mentioned earlier.  We do 10 microns of indium in this way, very
repeatably.

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Qinghong Du
Sent: Tuesday, 10 February 2009 3:47 AM
To: General MEMS discussion
Subject: [mems-talk] Bilayer PR for 2um Au lift-off process

Hi

I am looking for bilayer PR lithography to generate undercut for 2um Au
lift-off process. Anyone can sugguest some solutions? Currently I use
OCG825 and AZ5214E bilayer PR, but it is not very stable.

Thank you for your help.

Qinghong
DiCon Fiberoptics, Inc.