Hi, I have an issue with the image reversal process using AZ5214E. It's for
lift-off process. But we kept having troubles, e.g. part of the pattern
can't be lift-off or it took too long (more than 30mins in acetone with
ultrasound) to lift-off.

Finally we took SEM pictures and found out that the sidewall was a positive
tone like:

           ___
          /     \
_____/ PR  \_____

contrary to the negative tone which the image reversal process supposed to
give.

My main process is:

1. Dehydration at 150C for 10mins.
2. HMDS (I also tried processes without HMDS. No big changes.)
3. Spin AZ5214E @ 4000rmp 30s.
4. Prebake. 45s on the hot plate @ 110C.
5. Exposure with the mask for 5s. (Power ~ 8mW/cm2) Get about 40mJ/cm2. Soft
contact mode.
6. Post exposure bake (PEB) on hot plate @ 110C for 45s.
7. Flood exposure for 150s. It results in 1200mJ/cm2.
8. Develop in AZ422MIF for 2mins or 3mins. (usually I leave it in the
developer 1 more min after I can see the photoresist is washed out in the
huge pattern.)

The image did reverse and everything looks fine EXCEPT the sidewalls. The
photoresist didnot experience a temperature high enough to make it melt. It
looks to me somehow the top part was not well cross-linked and was finally
developed.

I also tried to change the prebake and, especially, the PEB bake conditions.
Then I have issues with over-cross-link.

Does anyone have similar experiences and could help me out? I'm kinda
exhausted. My plan is to base on the recipe where I got the over-crosslinked
pattern and tune down the temperature in the PEB bake. Am I in the right
direction? Is there anything else which needs to pay attention?

Futher details below:

One recipe I tried is from Univ. of Maryland. It prebakes 1min on the hot
plate @ 100C. Exposure: 40mJ/cm2. PEB bake for 45s @125C. Flood 1200mJ/cm2.
Then I had issues with the over-cross-link. The big patterns are OK but in
the fine pattern with 4 or even 8 microns gaps, I could see the photoresist
was not developed well there. And the pattern was bigger than the size it
should be. I tried PEB bake for 45s @ 108C. It was the same. And I'm not
sure about the sidewall. From the optical microscope, I suspect it's also
positive tone.

Another thing I tried is to replace the hot plate by the oven. I doubted
that the heat transferred from the substrate in the case of the hot plate
baking and maybe the top part of the photoresist could not experience the
same temperature and did not cross-link well.

I tried prebake in the oven for 30mins @ 90C. And 40mJ/cm2 exposure. For PEB
bake, I tried the oven @ 90C for 30mins, 10mins, 5mins, 2mins. Then flood
and develop. 30mins bake is definitely too much. I can see the pattern but
the photoresist was not washed clear away. in 10mins case, it is similar to
the case in the last paragraph. We could washed the photoresist away in
large gaps but not in the small ones. The situation happens in the 5mins
case too. However 2mins is too short and I couldn't get pattern. Futhermore,
with such short time (between 2min to 5min), repeatability will be a
problem.

****************************************************
* Zou Jie (Jay)
* Department of Physics
* University of Florida
* Tel: +1-352-846-8018
* Email: zoujiepku@gmail.com
****************************************************