Zhiyan,
    Using Ti and Cr in a Cl2/BCl3 plasma is not a good idea.  I use
Cl2/BCl3 to etch Cr and Ti metal layers using an RIE system on a
regular basis.  Chlorine-based plasma attack metals very actively.
    What may be happening is that your metal is delaminating because
the adhesion layer has been destroyed.
    You could try just depositing a layer of Ni and trying it without
an adhesion layer if you intend to expose it to Cl2 plasma.
-Eric


> ---------- Forwarded message ----------
> From: Zhiyan Liu 
> To: [email protected]
> Date: Mon, 22 Nov 2004 12:12:16 -0700
> Subject: RE: [mems-talk] Dry etch mask
> Dear friends:
>
> As suggested by MEMS discussion group, I tried to put Ti/Ni as a metal mask
> for ICP etching using BCl3/Cl2. And the thickness of the metal layer is
> Ti/Ni/Ti/Ni (50/700/50/700A). But right after i take the sample out of the
> evaporator, the metal layer starts peeling off due to the stress built up,
> or during the lift off, the metal is gone.
>
> I am wondering anyone has a suggestion for a good adhesion of this Ni layer
> on GaAs surface. I also learn that Cr can also serves as a metal mask for
> BCl3/Cl2. Does anyone use that before? and How is it compared to Ni?
>
> Thank you very much
>
> Zhiyan