That sounds like a textbook case of your PMMA getting unintentionally cross-
linked; high temperature or UV exposure could be a culprit.  You say that the
cracks appeared after evaporation.  Did you also inspect between the stripping
step and evaporation step?  Are you doing any post-develop baking, and if so,
what are the parameters?

I'm not personally familiar with that type of stripper, but the manufacturer
states that it heats the sample as part of its descum process.  What temperature
are you using in that step?  Is anyone else in your facility successfully using
that tool for lift-off descumming of PMMA, and if so, can you compare process
recipes?  As a test, you can always try your process without the descum step;
your metal adhesion won't be as good, but if the cross-linking goes away, you've
tracked down the problem.

It's possible the sample is getting too hot during the evaporation step, but
0.06 nm/s (0.6 Å/s) is a low enough rate that I'd be somewhat surprised if that
were the problem (unless your belljar is small so that the sample is relatively
near the crucible).  I used to evaporate up to 1.5 Å/s of Ti onto PMMA with no
problems, though the heating really depends on the details of your evaporator,
and the deposition rate which won't cause heating problems is pretty much
something you have to figure out empirically.



On Jun 1, 2010, at 23:08, landobasa wrote:

> Dear Researcher,
>
> I am still a beginner in fabrication, and I have these following problems in
the lift-off process. In order to provide the context, let me briefly write the
process steps:
>
> 1) PMMA film (200nm thickness) is patterned using Electron  Beam Lithography
(developed for 60s and IPA for 30s).
> 2) Patterned PMMA film was subjected to UV Ozone Stripper (SAMCO UV-1) for
60s.
> 3) The sample was deposited by Ti with 20nm thickness (with the rate of
0.06nm/s)
> 4) The sample was soaked in Acetone.
>
> Problems:
> 1) There are cracks everywhere in PMMA film after E-Beam Evaporation.
> 2) The PMMA cannot be removed by acetone.
>
> Any help would be very much appreciated.
>
> Best Regards,
>
> Lando