Sorry, for another mail, but....
The strength we have achieved is 5 GPa, not 5 MPa as reported below.

On Wed, 17 Apr 1996, Amit Lal wrote:
>
> Hello Hakan:
>
>   We have found that the strength of bulk micromachined parts can be
> significantly increased by a thermal oxidation. The thermal oxidation
> fills in any microscopic cracks that develop during dicing and
> polishing. A good reference on the increase of
> silicon strength due to thermal oxidation is:
>
> "Micromechanical fracture strength of silicon"
> Fredric Ericson and Jan-Ake Schweitz
> J. Appl. Physics. Vol. 68 No. 11. DEc. 1990. pp5840-5844.
>
>   Stress concentration and the related fracture failure at the V-grooves
> can be avoided by a plasma or isotropic etch of silicon to reduce the
> radius of curvature. A good paper is in the Transducers 1991 proceedings
> written by Kurt Peterson and others.
>
> With these two techniques we have been able to achieve 5 MPa fracture
> stresses in ultrasonically driven silicon structures.
>
> Amit Lal.
> Dept. Of. Electrical Engineeering.
> Berkeley Sensor and Actuator Center.
> University of California, Berkeley.
>
>
> On Mon, 15 Apr 1996, HaKAN ELDERSTIG -1080 wrote:
>
> > Silicon is a strong material they say....
> >
> > We have put our bulk micromachined silicon chips in a plastic
> > package. Some of the chips breaks or we can  se fine cracks, usually
> > close to the v-grooves. There is a bonded glass lid on top of the
> > v-grooves that should strengthen the device. The chips also had been
> > edge polished after dicing with 3 microns and 0.3 micron papers.  Do
> > anyone know if the polishing itself can lower the mechanical strength
> > of the chips? Or is the dicing bad enough?
> >
> > Hakan Elderstig
> >