Hi,
I am trying to make a PCR device using SU-8 2050 and SU-8 2002.  Here is my
process and my observations:

1.  Pattern 7-8um of AZ9260 (spin coat 2400rpm) on SiO2 wafer followed by 110'C
bake for 165s.

2.  E-beam evaporate 10nm Chromium followed by 600nm of Gold on the opened areas
in the resist.  Lift off with acetone.  The lift-off process is relatively easy
due to the thick resist.  When I profiled the metal lines, I saw large spikes at
the edge of the metal lines going up as high as 6-7um, while the centre of the
metal lines are 600-700nm in height.  These spikes cannot be removed by
ultrasonic action.

3.  Spin coat 4 layers of SU-8 2002 at 1000rpm (Microchem's datasheet reports
that at this speed, SU_8 2002 should coat to 3um each time).  After coating the
4 layers with soft bakes in between each coat, I noticed, to my suprise that the
total film thicknes was only 2-3um.  This reading is consistent as I have done
this multiplayer coating a few times before and each time, the thickness of 4
layers SU-8 2002 was only 2-3um at 2000rpm each time.   Can somebody help
explain this?

4.  As the spikes are not covered with this thin coat of SU-8 2002, I spin
coated SU-8 2050 at 6000rpm hopefully to obtain as thin a coat as possible
(<20um) on the metal lines.  The metal lines and spikes were satisfactorily
covered with the SU-8 2050, as verified by profilometer.

5.  Next I spin coat SU-8 2050 at 1500rpm in order to form the chamber wall for
the PCR device.  Previous attempts at this speed was able to obtain 110um tall
wall.  The coat after spinning was satisfactorily planar.  However after soft-
baking at 65'C for 5mins and then ramping up to 95'C for 30mins, I noticed
plenty of divots formed on the surface of the film during the 95'C soft bake.
Can somebody please help me understand this as this has not happened in my
previous attempt?

Thanks!

Jeffrey