kenney at slac.stanford.edu
Fri Dec 2 16:40:07 PST 2005
Request to use an aluminum holder similar to the one
used in the old stsetcher for the past five or more years.
Between 25 and 50 percent of the wafers etched in the old
machine use th holder.
Allowing a holder in the new machine would increase the number
of processes and people who could use it substantially.
1) Contact info
Chris Kenney and Jasmine hasi
Coral: kenney or jasmine
Phone: 725-3661 or 3655
Aluminum or aluminum oxide
Machined at Stanford
Many processes in the STS DRIE systems involve multi-hour
Given the lower loading near the wafer edge, the etch tends
to proceed faster at and near the wafer's outer radius.
When coating a wafer with resist, the few millimeters near the
wafer edge develop a standing wave pattern with local variations
in the film thickness. In particular at a radial distance about
2 mm from the edge there is a significant minimum in the resist
During long etches the resist in this local minimum tends to
be etched away and the silicon etched to significant depths.
This is especially true when the features being intentionally etched are
as this allows for most of the wafer thickness to be etched through
if a large area (~100 microns) is exposed near the edge.
Also the very side of a wafer, which is typically curved to
reduce stress, is etched quite badly during long etches.
Any damage near the edge of a wafer greatly increases its likelihood
of breaking either within the STS process chamber or in another
machine later in the process flow.
A second reason for a holder is that many etches make the
wafer more fragile. Most users feel more comfortable using
the holder in such cases as a safety feature incase the wafer
breaks in the process chamber. I have had wafers
break in the STSETCH1, but be caught by the holder and
unload normally without affecting other users.
Potential objections involve thermal issues and sputtering
of the aluminum.
Thermal issues will depend on the exact recipe and how it
is used. I believe thermal issues are best resolved by trial
and error by users as this will basically affect individual users
by damaging their resist and wafers.
The ion energies in the old and new STSETCHERS are very similar
given the essentially identical process pressures. Since there
is now significant problem with this in the old STSETCH
it is unlikely to be a problem in the new one.
Again a straightforward experiment can resolve this issue by looking
for excess grass when using the holder.
Indentical use as in the original STSETCH.
Potentially identical to the existing STSETCH holder.
Please ask if you have any questions about this.
Thanks for your time,
Chris and Jasmine
More information about the specmat