Michael,
             You were probably using older non reliable methods of defining the
lift off resist.  In my experience the best results to date have been 800
Angstrom lines and spaces with no problem with repeatability and processing.
Using image reversal the original definition is the defining step. Cray research
used lift off for 0.1 micron metal gates in production. The image reversal
creates the sloped side walls and the flood exposure the angle of the side wall.
Control from +22 degrees to -22 degrees.  A lot of users use 90 degrees for the
ultimate in resolution.  perfect vertical side walls.  Contact me if yoy have
any interest in technical papers.

 Bill Moffat
  CEO
 Yield Engineering Systems.
 408 954 8353.
[email protected]

________________________________

From: [email protected] on behalf of Michael Rust
Sent: Mon 3/5/2007 6:04 AM
To: General MEMS discussion
Subject: RE: [mems-talk] etch gold or lift-off?



>From my experience, liftoff becomes more challenging when your minimum
features and spacing become very small (increased pattern density).  For
instance, I found that trying to make 10 micron metal lines that are
separated by 10 microns spacing via liftoff was quite difficult.  When I did
the same design and used etching (actually, the mask would be the inverse of
the liftoff mask) I got much better results.  I'm not sure what kind of
pattern density you will have, and I believe there are ways to optimize the
liftoff process, but the quickest solution might be to do the etching.

To follow up on the comment from Nicolas, I also agree that using the
evaporator to deposit 1 micron of metal might be challenging.  As mentioned,
you may see cracks in your film or it might completely peel off (depends on
the metal and the substrate).