[mems-talk] Humps on SU-8 based micro-PCR chamber

[Yue Mun Pun, Jeffrey]

Hi,
I am fabricating a micro-PCR chamber using SU-8.  Here is my process below:
1.  Patterned Gold/Titanium heaters and thermometers with spacing of 100microns line/space on SiO2 wafer by evaporation.  The heaters are serpentine in shape and about 300nm thick.
 
2.  Spin coat 4 layers of SU-8 2002 at 3000rpm each layer on top of the heaters and soft-baked at 95'C for 3mins.  The total thickness of this SU-8 2002 layer is about 2.3 microns.  This layer is then patterned with 365nm UV (i-line) to a dose of 140mJ/cm2 and post-exposure baked.
 
3.  Spin coat 1 layer of SU-8 2150 at 3000rpm on the previous SU-8 layer and soft-baked at 65'C for 7mins and 95'C for 45mins.  This layer was aligned to the SU-8 2002 and exposed at 365nm to a dose of 260mJ/cm2.  Then the whole stack was post-exposure baked at 65'C for 5mins and 95'C for 15mins.  This SU-8 2150 layer forms a 150micron high wall surrounding the micro-PCR chamber.
 
4.  The whole stack was then developed in SU-8 developer for 15mins with agitation.
 
5.  After development the stack was hard-baked at 150'C for 15mins.
 
After the device was fabricated, inspecting under the microscope, I noticed that there were streaks of SU-8 formed on top of the SU-8 2002 floor.  I believe these streaks are the remains of the SU-8 2150 layer and they preferentially aligned themselves along the 100micron space between the heaters.  Examining these streaks with the surface profiler, I discovered that these streaks were actually SU-8 humps as tall as 30-40microns formed in the chamber itself on top of the SU-8 2002 layer, which is the floor of the micro-PCR chamber.  
These streaks were first noticed when I was soft-baking the SU-8 2150 cum SU-8 2002 stack, but at that time they appeared as notches at the edges of the device.
 
Has anyone noticed this interesting phenomenon before?  How do I get rid of these humps?  Are these humps formed due to the surface stress between the Gold conductors and the SU-8 2002 during heating and cooling?  If so, then I may need to use an adhesion layer such as Omnicoat?
 
Can anyone please advise.  Thanks!
 
Jeffrey