Hope I can partially answer you questions:

1. For the case of 500um thickness, I think either of the solution would
work fine. As far as I would concern, TMAH gives a little bit better surface
quality.

2.silicon oxide is easy to achieve good quality from thermal oxidation,
while nitride has bigger selectivity but you will often meet the problem of
pin holes unless you nitride deposition is optimized enough.

Good luck,

Hongjun Zeng, PhD
Microfabrication Application Laboratory(MAL)
University of Illinois at Chicago

-----Original Message-----
From: Simon Garcia
Subject: [mems-talk] Through wafer etching with KOH - mask

I have three questions about selectivity of wet etchants between silicon and
silicon oxide/nitride mask.

First, some background: A friend of mine wants to etch through a
0.5-mm-thick Si(001) wafer using KOH or TMAH. Both sides of the wafer have
2-micron-thick, SiO2. He wants to pattern the back of the wafer, and etch
holes all the way through to expose the SiO2 on the front from behind.
Without losing all the SiO2 on the front. I think he wants a few hundred
nanometers of the SiO2 left, but the more, the better.

My first question is: Where can I find information on the selectivity of KOH
and TMAH for different concentration and for different temperature, and
about adding isopropanol to the etchant?

I looked on the archives, and I see that with 30%, 80-deg KOH, I get SiO2:Si
selectivity of 100:1, and for 8% TMAH about 1000:1. I also saw a
recommendation for using 14% KOH to get the best selectivity.
Finally, I saw a reference to the Handbook of Microlithography,
Micromachining, and Microfabrication, vol 2, chapter by DL Kendall and RA
Shoultz. Do you all recommend this reference for useful information?
I will look for it.

My second question is: are there pros and cons between using silicon oxide
or silicon nitride as the etch mask?

My third question is: Is there any common tricks or pitfalls when etching
through a wafer using wet etching? I already know about the
(111) limiting planes. I don't think the shape of the holes is a problem,
but if (111) planes slow down etching too much, then does it work to use
(110) wafer?

Sorry for putting too many questions in one email! My feeling is that 5-8%
TMAH is best, but if anyone has advice or experience on this kind of
etching, I would be overwhelmingly grateful.