Hi John,

I've tried this and it worked quite well. I was electroplating within
patterned openings on 25-50um high strips of SU-8. I used a sputtered
seed layer consisting Cr (500-1000 Angstroms) / Cu (2000 Angstroms).
You get best (adhesion) results if you electroplate immediately after
depositing the seed layer. Remember to rinse the Cu seed surface layer
in 10% H2SO4, then thoroughly in DI water prior to electrodeposition,
if you cannot electroplate soon after seed deposition.

The only problem I faced was occasional edge peeling of the SU-8
strips, causing discontinuities in the seed, electrically isolating
some strips.

Surface roughening is unnecessary. If you are sputtering, and your Cr
seed has poor adhesion, try adding some O2 to the normal Ar plasma.
The resultant Cr2O3 formed the interface yields better adhesion to
polymer surfaces. Generally, though, sputtering Cr onto a polymer
surface under Ar plasma is sufficient to achieve good adhesion. I
tried both methods and there was no observable difference -
qualitatively, only.

Have a look at the my paper for further details on the process: M P
Larsson et al 2006 J. Micromech. Microeng. 16, pp.2021-2033

Good luck,

Michael


 > ---------- Forwarded message ----------
> From: "John Lee" 
> To: mems-talk@memsnet.org
> Date: Sun, 11 May 2008 18:41:02 +0300
> Subject: [mems-talk] Metal plating onto SU-8 after hard bake
> Hi,
>
> I have a problem regarding plating copper onto hard-baked SU-8 3050
> layer (SU-8 3000 series by MicroChem).
> So far, the following experiments have been conducted:
>
> 1. Electroplating of Cu on SU-8. Tried with/without surface treatments
> (i.e. plasma, mechnical roughening, permaganate). In all cases,
> adhesion was poor.
>
> 2. Sputtering (PVD) of 0.5 um Cu on SU-8. Very poor adhesion, failed
> in tape test (all copper was left on tape).
>
> 3. Sputtering of 0.5 um Cu on Ti adhesion layer on SU-8. Same result as 2.
>
> Does anyone here have experience of plating metal on hard-baked SU-8
> with good adhesion results?  If so, I would be glad to know the
> technique used and the obtained results as well.
>
>
> Thank You
>
> John Lee