I'm trying to create a regular array of 20 µm pillars of 40 µm height
in PDMS by casting against an SU-8 master.  In one design the pillars
are spaced 7.5 µm apart while in another they are spaced 25 µm apart.

So far we have been unable to create good molds.  The holes are
either: 1. ill-defined; 2. not full depth; or 3. tearing the PDMS
pillars off.

Thoughts so far are that the SU-8 was not properly developed, which
would cover 1 and 2 and that silanising the mold would enable easier
removal of the PDMS, hopefully fixing 3.  I would like some advice
first so that I can optimise my process sequence fully all at one go.
Is it better to over-expose the SU-8 slightly so that there is
definitely no under-cut which could 'grab' on the PDMS?  Is a long
(15-20 min.) development harmful to the mold?  Is the silanisation
permanent and if not how often would it have to be renewed (every
casting, every second)?

Also

We are using an MJB4 mask aligner for the SU-8 exposure.  We use
standard film photomasks.  The MJB4 has an uncollimated light source,
but I have been informed that the light source is well defined by the
internal optics so any difference from a collimated beam is minimal.
However I have also been informed that getting a well defined array of
holes in the SU-8 of the dimensions I want (or even dropping to 20 µm
depth so that the aspect ratio is 1:1) is extremely difficult and that
I should shift to silicon.  Having never worked with silicon this
would slow me down massively.

If anyone has an informed opinion of the possibility of producing a
mold of the dimensions I need it would be much appreciated.

If any further information is needed please don't hesitate to contact
me.

Thanks all

Matt

Dr Matt Davies
Post-doctoral Research Fellow
Science & Technology Research Institute
University of Hertfordshire
College Lane
HATFIELD        AL10 9AB
UK

e : m.davies@herts.ac.uk
t : +44 (0) 1707 28 61 74
f : +44 (0) 1707 28 1306