Two things need to be considered for a lift off process.

1. You need a point where you have a negative slope. The classic
solution is a chloroform soak of the spun resist to create a crusting and
so your resist profile has a lip.

2. The Evap process needs to have the poorest step coverage as possible.
Make sure your pressure isn't too high, or your dep rate not too high. A
high pressure causes scattering of the metal atoms on their way to the
wafer. You want line of sight from the metal vapor above the target.

The evap forms a metal vapor gas cloud immediately above the target. You
can cause ionization of the metal by the e-beam if the cloud is too
large or dense. The deposition goes form a gas phase to a line of sight
molecular deposition at some point in the drop off of the density.

Also a lower dep rate of metal atoms at the surface of resist gives more
time for the metal atoms to move to existing deposition than starting
new nucleation. This helps poor step coverage.

I would just do a Design of Experiment to see which condition works for
you.

Ed

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Le Cao Hoai Nam
Sent: Saturday, December 22, 2007 4:14 AM
To: General MEMS discussion
Subject: [mems-talk] Lift off process with S-1818

Dear MEMS reseachers,

I am using S-1818 photoresist for a lift off process of Au/Cr on
borosilicate glass substrate. The thickness of S-1818 is around 2um. And
the thickness of Au and Cr is 50nm and 80 nm respectively. After a Cr/Au
evaporation, I soaked the chip in an acetone container in an ultrasonic
bath. I noticed that even after 50min of ultrasonication the photoresist
still remained (the manufacturer recommends 10min of ultrasonication)
around 50% on the chip. On the other hand, it seems that I didn't have
the undercut pattern after development. I don't know if this have
anything to do with the remaining of photoresist. I would appreciate it
very much if someone could shed light on this problem.