NIk:

I will guess along with you with respect to formation of silicon nitride  afm
tips. I would assume a starting point of say 4-700 angstroms of silicon
dioxide on a prime silicon wafer followed by whatever thickness of silicon
nitride you eventually want to create somewhere around 1 micron I would think
would
be good. Using a positive photoresist with a pattern of circles to a
thickness of around 1 micron do a standard photo processing sequence with coat
bake
align develop. At this point you will want to reflow your photoresist  pattern
using a hot plate in a nitrogen box. This would work better if you can  pull
vacuum on the wafer to the hot plate. You will need to find out thru  testing
at what temperature you will get a true flow of the resist. With some  testing
you should be able to get a rounded feature somewhere around 50 degrees  from
vertical. Once completed you will need to develop a plasma etch process  that
will yield as close to 1:1 selectivity of the photoresist to the silicon
nitride. This might be accomplished using SF6 + Oxygen + Nitrogen in an ICP/RIE
type system. I would use SF6 because it will be more aggressive to silicon
nitride than say CF4. Once you determine the etch rates for photoresist and
silicon nitride you would just run the process until all of the photoresist is
gone. This should leave a nice sharp silicon nitride point for your AFM tips.

We have a plasma etch tool that I described above but there are engineering
costs and other charges required for use if you want to actually make some of
these devices. Good luck Bob Henderson