Andrea,

Many of the responses that have been posted have good advice.  As Edward
mentioned, 54˚ is extremely poor for a contact exposure, so it is very
reasonable to check out your fringe pattern with a clear mask.  You may not have
good contact, and adjustment may be required.  I had very robust processes with
Karl Suss aligners and typically was able to achieve <1µm features through 2µm
or more resist, but the tool required regular (at least weekly) maintenance to
keep the contact pressure consistent.

 Here are a couple of more suggestions:

1. Determine what the minimum exposure to clear the resist is for your
particular system (i.e., resist thickness, bake parameters, exposure tool/dose,
developer, etc.).  If you want to clear the resist in, for example, 45 seconds
of development, do some simple experiments to see what dose gets you there.
Once you determine that, I would use 1.5x to 2x the minimum exposure, and keep
the developer fixed at 45 sec.  The mantra I used to get near straight
sidewalls, was: over expose and under develop.

2. Have you determined that the 54˚ sidewall angle occurs after development
rather than after hard bake?  Perhaps the resist is flowing after development
during the hard bake (unlikely, but eliminate what you can, and you will be left
with the right answer).

3. Is there a difference between patterns in the center compared to the edge of
the wafer?  If so, this is a good indicator that contact issues are at play.

4. Is your mask clean?  Particles of resist stuck to the mask (a common problem
in contact litho) could be causing problems.

5. Do you have an edge-bead issue?  Is the photoresist building up at the edge
of the wafer, again preventing contact.

6. Although I have not used S1813 for a long time, the soft bake seems too long,
assuming that you are using a hot plate.  It is not likely, though, that this is
the cause of the sidewall angle problem.  However, if you are using an oven, it
is much too short (try 20 min.).  This would be a candidate because
insufficiently baked resist is less photosensitive than properly baked resist,
and in an oven the solvents are removed from the top to the bottom.

7. Check the dilution of your developer.  One that is too basic could cause
sidewall problems as it tends to erode the unexposed resist more quickly.  A
quick experiment would be to take an unexposed wafer and put it into developer
for several minutes.  You should see no more than a 5% - 10% thickness
reduction.

Brad Cantos
brad.cantos@holage.com
http://holage.com


On Jan 17, 2010, at 8:15 AM, Andrea Mazzolari wrote:

> Hi Prasanna,
> PR thickness is about 2 um.
>
> May you give me the recipe you used to obtain vertical sidewalls in S1813
> ? I think it could be a good starting point for me.
>
> Thanks.
> Andrea