I have faced the problem of peeling off of very thin structures (less than
15-20um) for a long time from SU8. I figured out that I was using mylar
transparency mask for exposure and somehow, it was the cause. When I used a
soda lime chrome mask for exposure, the features turned out great and there
was no peeling.

If you have really thin structures ( or low aspect ratio ) that are peeling
off and you are using a transparency mask, I suggest trying out a glass mask
( you can play with the contact mode first ).

Best,

-Abhishek


On Tue, Aug 5, 2008 at 6:30 AM, Gareth Jenkins
wrote:

> I've read HMDS doesn't always work for SU-8 but baking post HMDS
> spinning can improve the situation.
> (see this excellent website for more details:
> http://memscyclopedia.org/su8.html ).
> You should consider the following in your process:
> 1. Increasing your exposure dose. You generally need much more that
> the datasheet values for best adhesion.
> 2. Make sure you ramp up and down during PEB.
> 3. Do you really need the hard bake? This can actually increase the
> likelyhood of delamination.
>
>
> On Mon, Aug 4, 2008 at 19:43, Jose Guevarra  wrote:
> > I did try HMDS and that seems to work fairly well.
> >
> > Thanks.
> >
> > Oakes Garrett wrote:
> >>
> >> You should try the SU-8 adhesion promoters that are on the market.  MCC
> >> and Silicon Resources in Chandler Arizona both have products that work
> >> well.  What type of wafer are you using?