I can confirm that one can often get rid of the cracks with a final
hard bake (e.g. 2 hours, 135°C, in oven, not hotplate for 50 microns
thick SU8.).
This might be surprising, but it works. The explanation would be that
the SU8 have a glassy behavior at sufficiently high temperature. The
temperature might depend on the resist you use. Consequently, do not
let it in the oven too long, otherwise it could broaden your object.

On Fri, 27 Apr 2007 09:52:05 +0100
  Gareth Jenkins  wrote:
> At what point do the wrinkles occur? Possible causes are uneven
>hotplate temperature or too rapid ramping / cooling.
> Pitting can be due to insufficient dehydration of the wafer prior to
>spinning, contaminants (either on the wafer surface or within the
>SU-8) or bubbles during dispensing onto the wafer.
> As for the cracks, often these are too shallow to be a problem but
>are usually caused during the post exposure bake. Use a lower
>temperature bake and/or ensure rapid heating and cooling is avoided.
>Alternatively, I have also heard that a hard bake step sometimes
>closes these cracks.

**********************************************
Gwennou Coupier
Laboratoire de Spectrométrie Physique
Université Joseph Fourier
140, rue de la physique, 38402 St Martin d'Hères Cedex, France
Tel : 04 76 51 47 60
[email protected]
www.coupier.eu