Someone can correct me if I'm mistaken, but I don't think you'll ever
get PDMS to bond to Au in the same way it bonds to glass, ti or cr. When
you plasma activate PDMS it will bond to oxides, and since Au doesn't
oxidize significantly, you probably won't be able to bond to it. I
haven't tried it myself, but I imagine ITO (indium tin oxide) would be a
good material to use as an electrode for PDMS to bond to.
-Joe Grogan

N H wrote:
> At first: Thank you very much for your kind mail Mr. Kaplan.
>
>    As the matter of fact, I had done like your suggestion before
> sending my former message, and the sealing was very good, but the
> electrode was not active anymore. I tried to etch the Ti layer on
> active Au area, but PDMS did not work very well as "mask" and I had
> lost the control  over electro's active area again, by under-etching.
> By its turn, any solution involving alignment is forbidden, since in
> my process I have no chance to settle PDMS channels over the glass
> plate with that great precision.