SU-8 is extremely sensitive to radiation below 350 nm.  You should try a cutoff
filter which only allows radiation of 350 nm and above to pass through.  This
should change the sidewall profile and make it much more vertical.

Best Regards,
Garrett Oakes

EV Group
invent * innovate * implement
Litho Application Engineer - Direct: +1 (480) 305 2443, Main: +1 (480) 305 2400
Fax: +1 (480) 305 2401
Cell: +1 (480) 516 6724
E-Mail: G.Oakes@EVGroup.com, Web: www.EVGroup.com

-----Original Message-----
From: 郑瑞麟(Ruilin Zheng) [mailto:zhengruilin@gmail.com]
Sent: Sunday, May 04, 2008 5:48 PM
To: General MEMS discussion
Subject: Re: [mems-talk] SU-8 partially concave sidewall profile problem

well, thank you all at first.

It works as I doubled the exposure dose. But the inclined sidewall profile
still show up, but less severe, the slope is about 8/390, and still the
lower part close to wafer appears more inclined than the upper part near top
surface. The UV light source I am using is Karl Suss uv400, 10.7mJ/cm2 at
405nm, 4.9mJ at365nm, and a short peak at 334nm.

I also found a problem for the sidewall surface, it is corrugated. Does this
problem have something to do with internal stress?