For vertical sidewalls and wet etching, you're talking (110) silicon. One of
the more knowledgeable people in this area is D L Kendall, who published a
number of papers in the 1970s and 1980s and a few in the 1990s. Lots of
practical advice can be found in his papers. He also did a chapter in an
SPIE book, "Handbook of Microlithography, Micromachinining, and
Microfabrication, Volume 2".

Making a through wafer trench that is 2um wide may prove to be exceedingly
challenging. You'll probably run into diffusion limitations. Separate
effects may tend to widen the trench, such as a finite (111) etch rate,
slight crystal misalignment, and etch mask erosion/stress.

Good luck,

Bill Eaton, Ph.D.
Materials & Analysis Manager
NP Photonics

>
> Dear members:
> I have a question about silicon trench wet etching. If etched
> from back side,
> for example, 2um mask dimension, how wide the trench will be
> in the front side
> with wet etching? I know this depends on the orientation of
> wafer and the
> trench direction. According to the book, if the trench is
> aligned accurately,
> it is possible to get a vertical side wall, thus, the width
> in front side will
> be the same as that in back side, say, 2um.
>
> I donot know whether it requires very high accuracy to
> alignment in order to
> get vertical side wall. If I do the alignment manualy,
> generally how thin i
> can get in the front side? Is there anybody has experience in
> this area?
>
> thanks.
>
> Regards,
> Yilei Zhang
>
>