A 3~5 wt% addition of Ta to a gold melt resolves the spitting issue.  I'm not
sure how compatible Ta and Al will be, but it might be worth a try.

Dan

-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Brad Cantos
Sent: Wednesday, April 01, 2009 12:12 PM
To: General MEMS discussion
Cc: li shifeng
Subject: Re: [mems-talk] Pin Holes on the Al Surface

Hi Shifeng,

When you say "black stuff," are you viewing it with an optical
microscope?  If so, you may also want to look at it with an SEM if you
have one available.  What you describe sounds very much like spitting
during aluminum deposition with an ebeam evaporator.  This will cause
relative large spheres of Al to be forcefully ejected from the melt
and deposited on the wafer surface.  When they adhere to the surface,
they will block further deposition in that area by shadowing the metal
vapor.  These spheres do not adhere well and will drop off during
subsequent process steps leaving a pinhole that often appears black
optically.  One thing you can do is reduce the evaporation rate,  In
my experience 5-10Å/sec will eliminate most of this type of spitting.
There is a paper from IBM back in the 1970's that describe the
phenomenon and solve it by adding a small amount of another metal
(maybe Ta or W - I can't remember) to the melt.  I'll try to look up
the reference for you if you need it.

Brad Cantos
[email protected]