This could be a wafer prep issue.

Did you oxidise the wafer and thoroughly dehydrate? Oxygen plasma should do
both or use piranha (or conc. H2SO4) followed by 250C bake.

On 15 June 2011 00:35, Shane GUO  wrote:

> Hello,
>
> The pulling back did not happen in softbake but before it after a long
> relaxation. I used a rectangle silicon substrate. Usually the SU-8 retreats
> most from the four edges of the substrate but extends and covers the
> corners. The shape of the final SU-8 looks like a puddle surrounded by four
> concave lines. The flatness is not that bad for the SU-8 puddle.
>
>
> Cheers