Hi Heiko,

It could be the native oxide on the silicon wafer surface, formed
prior to the deposition of Cr/Au, that is etching and releasing the
layer. This is a possibility worth looking into as your etch step
involves BOX layer removal, meaning your sample will be in etchant for
a number of hours. The metal layer will also be under moderate tensile
stress, despite the annealing step, assisting film release.

I would perform an HF rinse prior to depositing the metallised layer.
The advice given by other respondents about diffusion barriers should
also be implemented. Try Ni or Pd as the barrier layer (i.e. Cr/Pd/Au
or Cr/Ni/Au). If you have a Pt target, you could also try depositing
this after Cr as well.

Good luck!

Michael

> Heiko Prüßner wrote:
> > Dear all,
> >
> > I had a 20nmCr/ 200nmAu film sputtert on silicon. This setup has been
treated at 350°C for an hour. At the final Hf etching step which I used to
remove the BOX of an SOI wafer the Au film peeled of.
> >
> > Does anybody have an explanation? Do I need a diffusion barrier between the
Cr and the Au? Or has anyone a suggestion for another adhesion layer?
> >