Hi Dhananjay,

Bill is right, you should try vapour priming in the forst instance.

If problems with delamination persist, try locking your SU-8 to the
substrate. I have developed processes to do this on Si and glass
substrates, and the results are very reproducible. I was focussing on
microcantilevers, and not microchannels, but the process can be
adapted for your application. I can also advise if you need further
guidance.

Journal of Micromechanics and Microengeering, Vol 15, 2005, pp.
2074-2082: for Si substrates

Journal of Micromechanics and Microengineering, Vol 16, 2006 , pp.
S161-S168(1): for glass substrates

Let me know if you would prefer me to send the PDFs.

Regards,

Michael Larsson

>
> ---------- Forwarded message ----------
> From: "Dhananjay Dendukuri" 
> To: mems-talk@memsnet.org
> Date: Thu, 23 Oct 2008 11:48:30 +0530
> Subject: [mems-talk] SU-8 delamination problem
> Dear all,
>
> I have a severe problem of SU8 delamination. Please go through my process
> below and advise
>
> I am fabricating an array of 100 microchannels of 50 micron width and 40
> micron depth (the gap between two channels is 100 micron) on a 4 inch
> silicon wafer using *Su8 2035* negative resist. The process flow is given
> below.
>
> 1. Hard bake - 150 C for 20 min (after cleaning silicon wafer in Piranha
> solution).
> 2. Spin coat - 15 sec at 400 rpm followed by 30 sec at 2000 rpm.
> 3. Soft bake - 3 min at 65 C and 5 min at 95 C. (wait for 10 min)
> 4. UV expose - 28 sec to get an exposure energy of 220 mJ/cm2.
> 5. PEB - 30 min at lower initial temperature of 65 C and ramping up the
> temperature to 95 C in every 15 min with every 10 C increase. (Wait for 10
> min)
> 6. Development - In ultrasonic bath (without sonicator SU8 in the channel
> path is not possible to remove completely).
>
> Within 1 min of development, all the channels are completely delaminated. We
> tried the process on many wafers by changing different soft bake and PEB
> timings, but nothing is working.  Any suggestions to avoid delamination
> problem are deeply appreciated.
>
> Thank you
>
> Dhananjay