SU-8 adhesion problem

Mary Tang mtang at
Fri Aug 31 17:04:57 PDT 2007

Hi Zubin --

Normally, SU-8 adheres very well to SiO2.  Two things you might consider:

1.  Make sure your exposure is long enough, so that the SU8 at the 
substrate is fully cured.  For very thick SU8 (though your 10 microns is 
not very thick), some people use an optical filter (lay it on top of the 
mask) to filter out sub-i-line wavelengths to improve curing down to the 
substrate surface.

2.  Glass quality and the method by which the glass is formed can make a 
big difference.  Glass has a lot of sodium and may also have boron, Mg 
and Zn.  It is often not very uniform in composition and "dopant" 
segregation to surfaces is especially if the glass is cast or rolled 
from a nearly molten state.  High concentrations of sodium can lead to 
adhesion problems.  Slight etching in BOE or other HF can remove areas 
of high concentration (although this might make poorer quality glasses 
hazy.)  Etching also roughens up the surface a little which helps in 

Good luck!


Mary X. Tang, Ph.D.
Stanford Nanofabrication Facility
CIS Room 136, Mail Code 4070
Stanford, CA  94305
mtang at

Zubin Huang wrote:
> Hi,
> I'm having some problems of SU8 adhesion and hope someone could offer 
> advice.
> My substrate is 25mm*50mm VWR cover glass with thickness about 0.13mm 
> to 0.16mm.
> Following is my process:
> 1 Clean the cover glass with Aceton, Methanal, IPA 5min each in 
> ultrasonic bath.
> 2 O2 plasma treating for 10min in MRC (20sccm O2, 100mTorr, 50W)
> 3 Yes Oven for singe and HMDS treatment.
> 4 Coat 30s 3500rpm SU8 2010 at headway with 10um thick.
> 5 Bake 4min @ 95C
> 6 Exposure 4.5s *3 cycle with 30s spacing at Karlsuss;
> 7 Bake 5min @ 95C
> 8 Dev 3min immersion.
> 9 IPA rinse for 10s and blow dry.
> After the process I found the small features of 10um size on my 
> structure came off and changed position on the substrate. I would like 
> to have advice on improving the adhesion between the SU-8 and the 
> glass substrate.
> Thanks,
> Zubin

More information about the labmembers mailing list