Albert,

After spending years on bulk silicon etching in TMAH, KOH, and EDP I agree with
your statements. Oxygen concentration as it relates to oxygen precipitates in
your specific process are a major issue in the control of bulk etching. At my
former company we picked the starting material based on the dimensional control
spec. One aspect of feature control in bulk etching to keep in mind is the etch
rate of the planes that create your structure. For example, the 111 etch rate is
different for different etchants at various concentrations, temperatures, and
with different additives. Using an etchant with a higher 111 etch rate when
making a square cavity in a 100 wafer will uncover more defects in the 111
sidewall and will result in less dimensional control. We also had to consider
the dopant type and concentration in respect to the yield strength of membranes.
There is a lot of knowledge of the effect of the starting material on bulk
etching at some of the major silicon suppliers to the MEMS companies. There are
also low oxygen content wafers available that can result in better dimensional
control if you have a temperature history that produces precipitates.

I could have used some more available research on the topic when I was trying to
meet a spec.

Dan Chilcott
Sr. Member of the Technical Staff
ITT
dan.chilcott@itt.com

-----Original Message-----
From: mems-talk-bounces@memsnet.org [mailto:mems-talk-bounces@memsnet.org] On
Behalf Of Albert Henning
Sent: Monday, April 27, 2009 2:57 PM
To: mems-talk@memsnet.org
Subject: [mems-talk] silicon anisotropic etching vs. starting material?

Hello,

Over the years, I have observed (and this is purely anecdotal) that different
silicon starting material etches differently in, say, KOH or TMAH of fixed
temperature and concentration.  I have also observed that the yield strength of
silicon membranes, created using KOH or TMAH etching, depends upon the starting
material.

My conjecture on this observation centers on oxygen precipitates:  that higher
O2 concentration in the silicon leads to faster etch rate of higher-order
planes, and also to lower yield strength.

Microelectronic silicon technology benefited from decades of fundamental
materials work by NIST, and by individual researchers, to establish the
relationship between dopants and other silicon defects, and electronic
properties of Si.  MEMS has not benefited from the same intensity of research,
it seems to me, in terms of coupling dopant and defect concentrations to
mechanical properties of Si.

If anyone has experience or references on this topic to share with the
community, I'm sure they would be much appreciated, by me and by others.

Albert K. Henning, PhD