Hi Konstantin,

I've tried mechanical interlocking to successfully retain SU-8 to
Si/SiO2 substrates in KOH for several hrs (>10hrs at least). As for
minimising undercut rate, this is not assured as delamination of the
layer will occur where there are no interlocking lobes. You may be
able to slow the undercut rate, but it won't be reproducible.
Basically, chemical adhesion promoters will not help to retain
parylene-substrate adhesion at all in aggressive chemical media such
as KOH - it is very much a red herring. If your layer is delaminating
too quickly, then I would switch to a metallised layer (Cr for
example, whose oxide - Cr2O3 - forms a stable etch-stop. There is no
point in using a metallised layer beneath the parylene mask, as this
will only degrade feature resolution with no additional benefit. If
you have to use parylene try mechanical interlocking if your
application permits, if not, then try a metallised layer in isolation.

p.s. in Si, you need SiO2 at the interface to prevent degradation of
interlocked lobes. The duration of attachment is, therefore, some
function of the etch rate of SiO2 in KOH at the tempurature you set.
So, if you can apply an SiO2 layer in your application, why then
bother with the parylene? The papers below describe my efforts with
the attachment of SU-8 strips in KOH. The intention was not to mask
the underlying, but to retain features and allow the development of
polymer-Si MEMS hybrid structures.

[1] Journal of Micromechanics and Microengineering, Volume 15, Issue
11, pp. 2074-2082 (2005) - mechanical interlocking in Si substrates
(requires RIE)

[2] Journal of Micromechanics and Microengineering, Volume 16, Number
6, June 2006 , pp. S161-S168(1) - mechanical interlocking in generic
substrates (does not require RIE)

Good luck,

Michael Larsson

> -----Original Message-----
> From: mems-talk-bounces@memsnet.org [mailto:mems-talk-bounces@memsnet.org] On
Behalf Of Konstantin Glukh
> Sent: Tuesday, October 07, 2008 4:30 PM
> To: General MEMS discussion
> Subject: [mems-talk] Parylene mask for KOH etch
>
> Can anyone comment on using Parylene as a mask for KOH etch, either by
> itself or in addition to SiO2 hardmask? Is there any way to ensure
> good adhesion to Si and SiO2 and minimize undercut rate?