I determined the necessary soft bake time using the method of removing
the wafer during the bake and allowing it to cool, then placing it back
on the hotplate. Wrinkles appearing as it passes the Tg indicate
insufficient baking. I then repeated this until no wrinkles occurs. This
gave me a 2 hour bake time - twice as long as suggested in the datasheet
for a 260um layer. Hence I am fairly sure the soft-bake is long enough.
I now think it may be a levelling issue as the thickness of my
structures varies across the substrate from about 330um to 500um.
Overall it still seems too thick but maybe there is some additional
effect (perhaps some reflow from the edge bead region?).
My hotplate appears level using a bubble spirit level but I guess this
may not be accurate enough for thick layers which are more sensitive to
this. Any ideas on how to get really good levelling?

Gareth

Ali Bhagat wrote:
> I think another reason why you may be obtaining films thicker than
> what you
> expect is probably due to the fact that your soft-bake is not long enough
> and may be the films are still retaining quite a bit of solvent.