I want to make some 50um deep PDMS channels on a glass wafer surface.

However I can't use the typical PDMS process, which includes making master using
SU-8, then pouring PDMS and bonding PDMS with Glass.

Are there any suggestions?

My Idea:

1. Pattern the channel using negative mask (dark in channel area), use positive
PR (AZ4620)

2. Spin coat PDMS on top of AZ4620 pattern.

3. Spin coat AZ4210 or AZ4110 on top of PDMS

4. Pattern the AZ4210 using positive mask (transparent in channel area)

5. After exposing the channel area from last step, Use RIE to etch the PDMS away
and expose the AZ4620.

6. Use Acetone to remove the AZ4620.

My concern is the PDMS might have a climbing effect on step 2, so the PDMS
surface will not be flat.

Thanks for your help!

Best Regards

Ezone