Hi,

I have a MEMS vendor that has changed their anodic bond process to a triple
stack method. My problem is that we sputter metalize the back glass for
eutectic soldering with Ti/Pt/Au metal stack and when the wafer is sawn the
gold peels back from the edges and our requirements are no peeling at the
edges. The bond stack is glass/Si/glass

I believe this to be caused by the Anodic bond sodium concentrations built
up because the chuck is metal. I have researched and know about graphite
chucks and using a second glass wafer but I have wafers with what are
believed to be sodium and sodium hydroxide that prevent proper adhesion of
the base metal.

I currently plasma etch with Argon for 5 minutes and then with Oxygen for 5
minutes before sputtering the metal. Recently however this does not seem to
work after the vendor bonding method changed.

How can I inertly remove the sodium on the surface. The vendor says DI water
cannot be used with the MEMS and the exposed cavities on the side we have
the contamination on.

Can I plasma etch longer with Argon or Oxygen, which is better that will
actually remove surface material and the sodium. The sodium is visually seen
as white crates and spots or blotches on the backside. An EDX report shows
about 4.1% on the surface below the peeled metal.

I need to know how to resolve the metal adhesion problem when sodium is
present on the surface after anodic bonding.

Roger Horton
rhorton@s3cinc.com