Hello Paul,

I believe that your photoresist might go during DRIE through process similar
to DUV hardening so the chemistry changes and it is inert to developer.

I did in past the lithography in the cavities 250 um deep and 100 um wide.
That worked quite fine and I was able to pattern both on surface as well as
in cavity. The main obstacle was that with bath or puddle development the
resist did not develop within the cavity. I believe that either developer
did not get to the bottom due to surface tension or that refresh rate of
developer within the cavity was too low. Nevertheless, after using spray
development the result were fine.

We also wrote a paper about that
http://www.future-fab.com/documents.asp?d_ID=4719

Hope that helps.

With best regards,

Daniel Figura

smartfabgroupT Company
process consulting . data processing . fab software
Phone: +44 20 3286 4342
E-mail: daniel.figura@smartfabgroup.com, Web: www.smartfabgroup.com


-----Original Message-----
From: mems-talk-bounces+daniel.figura=smartfabgroup.com@memsnet.org
[mailto:mems-talk-bounces+daniel.figura=smartfabgroup.com@memsnet.org] On
Behalf Of Dibyadeep Paul
Sent: Tuesday, November 02, 2010 9:44
To: mems-talk@memsnet.org
Subject: [mems-talk] lithography on a wafer surface containing deep trenches

Hi,

I have a wafer, which contains some deep trenches. I fabricated these 250
micron deep trenches using DRIE, with AZ9260 as the mask layer. Now I have
to do a DRIE, to etch away 150 micron more, in the trenches already
fabricated as well as in some other regions.

In order to do this 2 level DRIE, I had grown an oxide layer 2 microns thick
to act as the masking layer for the second DRIE. I had planned to use 9260
as the masking layer for the first DRIE and the oxide as the masking layer
for the second DRIE. However after growing the oxide layer, and etching away
the unwanted regions, I found that some of the masking oxide layer has also
been removed.

As I said for the first DRIE, I used 9260, and the DRIE went fine. I had
initially thought that I would use the 9260 layer which would remain after
the DRIE and do a contact exposure-development, to expose the areas where I
wanted the next DRIE to happen. This would help me do the DRIE on regions I
wanted to expose, to circumvent the problem created by the presence of the
holes in oxide layer.

However after the first DRIE I found that the 9260 layer has become complete
inert to the the developer AZ 400K, even after multiple and long expsures.

Next, I removed the 9260 on one of the test wafers, and tried doing a
contact exposure-development with various photoresists in AZ and SC series.
In all of them I found that even after long exposures, some of the
photoresist always remains within the trenches, and I cannot remove that. I
am not sure why this happens.

So I wonder if somebody could help me with these questions:

a) how do I do a fabricate a photoresist based masking layer for a wafer,
which already contains deep trenches of ~ 250 microns?

b) Is it possible to activate the photoresist top surface after a DRIE, so
that I can do a contact exposure-development on the photoresist already
present?

Thanks

Dibyadeep Paul