Zhiyan,

I am assuming that the SU-8 is being lost by lifting off (i.e., the
patterns still exist, but float away in the bath).  If this is the case,
adhesion may be one potential problem (although, as you say, the
OmniCoat is supposed to improve adhesion to metal) or it may actually be
a process issue.

One thing to remember, first of all, is that the Omnicoat is intended as
a lift-off material for the SU-8, not as an adhesion promoter. How are
you developing the OmniCoat to allow access to the copper?  Be certain
that you are not over-developing.

Second possibility is that you are actually under-crosslinking the SU-8
through under-exposure.  This will typically present itself in a
lift-off manner, since the upper portions of the SU-8 become exposed,
while the lower layers do not receive sufficient dose, and dissolve
during develop. Try increasing your exposure (remember, the exposure
curves are based on silicon exposure, but you are on copper, with a thin
layer underneath that may be, depending on thickness, acting as an ARC).

Best Regards,
Chad Brubaker

 -----Original Message-----
From: Zhiyan Liu
Subject: [mems-talk] Adhesion problem of SU-8 on metal surface

I am trying to pattern SU-8 on the Cu seed layer and then do electroplating
of Cu. I applied Omnicoat, which can improve the adhesion between SU-8 and
metal layer according MicroChem, beneath the SU-8. But right during the
development, some of the SU-8 is gone. Then in the following electroplating
process, almost all the SU-8 is gone.

Does anyone can give me a hint on the improvement of SU-8 on metal such as
Au or Cu?