Hi Andrea,

The SOI is UNIBOND SOI from SOITEC, which uses the smart-cut process
instead of the SIMOX process. The thickness tolerance on any other kind
of SOI other than SIMOX or UNIBOND wasn't sufficient for my application,
and I didn't come across any vendors of SIMOX who could supply custom
layer thicknesses in low volume.

Jason Milne
The University of Western Australia
Microelectronics Research Group



-----Original Message-----
From: mems-talk-bounces@memsnet.org
[mailto:mems-talk-bounces@memsnet.org] On Behalf Of Andrea Mazzolari
Sent: Tuesday, 16 December 2008 2:18 AM
To: General MEMS discussion
Subject: Re: [mems-talk] Stress reduction in SOI

Hi Jason,

> I purchased some thin SOI recently to make some fixed-fixed beams and
> cantilevers, and I was assuming that the device layer would be close
to
> zero stress (slightly tensile, if anything) and have no stress
> gradients. Having made the structures, it seems that the device layer
is
> compressively stressed (the fixed-fixed beams buckle) and has a stress
> gradient (the cantilevers curl upwards). I was wondering if anyone has
> tried to anneal SOI wafers to reduce stress and/or stress gradients,
and
> if so, what conditions did you use?

Which kind of SOI are you using ?
For my experience if you use SIMOX SOI you should not have this kind of
problem.

Andrea