Patrick-

At the very worst the stress in the oxide will be about -300 MPa
(compressive) for a high quality thermal oxide.  Your case will depend
on the type of SOI, i.e. bonded-and-etch-back SOI, implanted oxide, epi,
etc.

The -300 MPa value can be calculated theoretically (and has been
confirmed experimentally) by noting the amount of expansion crystal
silicon will incur after each atom in the lattice incorporates two
oxides, and the molecular spacing for quartz.

As for how much the oxide will expand, the problem is pretty simple for
unbounded single material structures under internal stress, and can be
obtained by knowing the elastic constants (Young's modulus and Poisson's
ratio for isotropic materials).  For a laminate/composite, the general
case requires FEM, and for silicon, if you need the accuracy, you will
need anisotropic elastic constants.  These can be found on the web.

Raj


Patrick Ping-Chi Lu wrote:

> Hi,
>
> After releasing the substrate for my devices, which are made using SOI
> wafers, the sacrificial oxide layer wants to expand to its "natural size"
> since it was previously held under compressive stress (came that way from
> the vendor). I am trying to quantify the amount of compressive stress in my
> SOI wafers. I also want to know the amount that the oxide layer will expand
> when released. Does anyone have good tips as to where I can find this
> information, such as journal articles or websites? Any help would be greatly
> appreciated!
>
> Thank you!
> Patrick Lu

--

Raj Gupta, Partner
   Volant Technologies
   Nanotechnology & MEMS Consulting
   33 Vandewater Street, Suite 205
   San Francisco, CA  94133-1967
   OFF:  +1 415 362 3694
   FAX:  +1 253 563 1815
Website: http://mindspring.com/~gupta