Long Anneal Request

Chris Kenney drckenney at gmail.com
Thu Feb 1 14:45:44 PST 2007

Dear Specmat,

Long Anneal at 1100C Request

Chris Kenney and Jasmine Hasi
Coral login: kenney  and jasmine
725-3661 or 3655
Molecular Biology Consortium and University of Manchester

Material: ultra-high purity, float-zone silicon fresh
from the manufacturer, which is Topsil in Denmark.

website is:   http://www.topsil.com/1

Motivation: It is well established in the literature
that including oxygen in float-zone silicon improves
the radiation tolerance. Specifically the bulk displacement
damage caused by hadrons. See attached papers for details.

Silicon sensors will be used as close to the primary particle
beams as possible at the Large Hadron Collider, which is nearing
completion. These sensors will be exposed to fluences equivalent to
between 1e15 and 1e16 neutrons per cm2. This will result in
lattice defects, which act like acceptors and effectively
increase the p-type dopant concentration making the sensors
more and more difficult to deplete. Depending on the
sensor design, silicon is unable to operate beyond a set
fluence. Diffuing oxygen into the crystal allows the sensors
to operate after high fluences and therefore can substantially
extend their useful life.

This allows more science data to be taken.

Process flow:
1) remove pristine, float-zone wafers from the original cassette box
2) clean at wbnonmetal
3) clean at wbdiff
4) grow thin oxide (500A) as scarificial 'cleaning' oxide
5) strip oxide
6) grow oxide (5000A) to serve as oxygen dopant source
7) perform long anneal

In the literature this anneal is usually done at 1150C for
about 48 hours. When diffusing from both surfaces of a 300 micron
thick wafer, this results in a fairly uniform oxygen profile
throughout the wafer thickness. The literature indicates that
a oxygen level over 1e17/cm3 is necessary to gain the full benefits
of this approach.

Oxygen diffusion has a temperature dependence such that it roughly
doubles every 50C.

So our choices seem to be:
2 days at 1150C
4 days at 1100C
8 days at 1050C

Since our electron and holes have to drift across the complete
thickness of the wafer (i.e. 300 mircons), we require a high
bulk lifetime. We also need to keep the leakage currents low
as this increases the noise and can saturate the preamps.

Contamination is a large concern for us.

In the 1990s we often performed 16 hour anneals at 1150C, so
we hope that 1100C would not harm the tube glassware.

To avoid possible contamination, we would like to leave the
wafers in the tube continuously during the anneal.

Nevertheless, subdividing the anneal into a modest number
of high-temperature runs interspersed with cool-down periods
is fine.

So it would be very interesting and potentially very valuable
for us to do this long anneal.

That said, if there is a non-neglible risk that this would
damage the quartzware, we can skip this step.

Thanks for considering this,

Chris and Jasmine
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://snf.stanford.edu/pipermail/specmat/attachments/20070201/84e9b16f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ATLASoxy.pdf
Type: application/pdf
Size: 258659 bytes
Desc: not available
URL: <http://snf.stanford.edu/pipermail/specmat/attachments/20070201/84e9b16f/attachment.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rosetn00-01.pdf
Type: application/pdf
Size: 97606 bytes
Desc: not available
URL: <http://snf.stanford.edu/pipermail/specmat/attachments/20070201/84e9b16f/attachment-0001.pdf>

More information about the specmat mailing list