I always first deposit a very thin layer of SU-8 with no pattern on the
glass wafer as adhesion layer. Then go for thick layer SU-8 with the desired
pattern. Thinner layer of SU-8 has less stress due to the difference of
coefficient of thermal expansion.

Yifan Wu

On Fri, Dec 10, 2010 at 9:38 PM, Mikael Evander wrote:

> Hi all!
>
> I was wondering if anyone has some kind of input regarding how much better
> the adhesion of SU-8 3000 is compared to the 2000 formulation. I'm trying to
> make a 10- 20 um wide, 5 um high and up to 10 mm long "barrier"on a glass
> wafer. I did my first tests with SU-8 2005 this week and noticed that the
> adhesion was terrible. Very few of the barriers remained intact and after
> dicing the wafers most of them were gone. My wafers were piranha washed and
> HMDS-treated just before spinning the SU-8.
>
> I've been reading a bit and people usually recommend making sure that the
> exposure time is enough, doing a low temperature PEB, ramping of the hot
> plates, having a very clean and dehydrated wafer. Something that also comes
> up a lot is to use SU-8 3000 instead as it is supposed to have increased
> adhesion. My question is if anyone can comment on how well the 3000
> formulation seems to adhere to glass substrates. I need the structure to
> remain intact as it will be a mechanical part of a microfluidic device.
>
> Many thanks and have a great weekend!
>
>
> /Mikael