Hi Edward,

Thanks for all the useful information.

1. Pr thickness is 1um.

2. resolution is 500um.

Best regards,
Andrea

-----Original Message-----
From: Edward Sebesta [mailto:esebesta@tx.rr.com]
Sent: domenica 2 agosto 2009 21.49
To: mazzolari@fe.infn.it; 'General MEMS discussion'
Subject: RE: [mems-talk] s1813 question

There are several issues here.

1. How thick is your resist coating? Do you have topology on your wafer
and to what magnitude? What resolutions do you want to achieve?

2. You will need at least 30 minutes and maybe as long as 45 minutes to
bake them. In a convention oven the boats don't reach bake temperature
for 20 to 30 or more minutes typically, and you will want them to bake
for some time after the oven and boat have reached the target
temperature.

3. If you have one or two microns of resist and about a 1 micron of
topology and you aren't interested in driving resolution and process
repeatibility you could bake them 30 or 45 minutes at 100 degrees or 110
degrees. Probably if you bakes 45 minutes at 100 or 110 degrees it would
work for high topology wafers.

4. Convection ovens are horrible for process control if you are pushing
resolution. However, adopt these rules, one boat at a time, don't open
the door when it is baking, and put the boat in the same location in the
oven each time. It will help some, but not that much.

5. You can spin resist on both sides of a wafer the following wafers.

        A. Index your wafers so they move through with a set orientation
down the spin track. The wafers can be spun with edge contact. The hot
plate could contact don't care areas. Or perhaps some support that is
edge support.

      B. Use soft rubber suction cups on the chuck to spin the wafers.
They won't tear the resist. Make sure your oven uses three pins to
support the wafer during the softbake and take the yield hit.

6. If you wish to drive your process to greater performance, do a Design
of Experiments and execute it. In this case one of the factors will be
to have robustness against convection oven variation.

Ed Sebesta
Semiconductor and MEMS process engineer.