[mems-talk] Isotropic silicon etch

[Kirt Williams]

Dan--

HF-nitric-water or HNA silicon isotropic wet etchants may work, but are 
notoriously nonuniform due to loading effects.
Nonuniformity is greatly reduced by agitation. They also tend to have lower 
etch rate with use, leading to poor repeatability.

I would expect an SF6 plasma to be be more repeatable, because you always 
have fresh etchant (F radicals) being formed in the plasma.
Plasma etches tend to etch faster at the outside of the wafer due to less 
loading there.
I'd try a pressure at the lower end of your system's range for the best 
uniformity.
I'd try a low power (say 50 or 100 W) for more isotropic character.

Another option is XeF2 plasmaless isotropic silicon etching. With the 
homemade systems I have used, this is extremely nonuniform.
Xactix, a commercial vendor, may have this problem licked.

    --Kirt Williams

----- Original Message ----- 
From: "Chilcott, Dan W" <Dan.W.Chilcott at delphi.com>
To: <mems-talk at memsnet.org>
Sent: Wednesday, June 27, 2007 11:02 AM
Subject: [mems-talk] Isotropic silicon etch


To All,

I am looking for a well controlled isotropic silicon etch for etching 2 um 
deep features. I do not believe that a timed plasma etch would be controlled 
enough for this application. I am looking for a control of better than 0.5 
um from lot to lot and across the wafer. Any ideas?