---------- Forwarded message ----------
Date: Tue, 24 Sep 2002 13:13:36 -0700
From: Chris Wrede <wrede@triumf.ca>
To: Dave Ottewell <dfot@triumf.ca>, dauria@sfu.ca, hussein@unbc.ca,
     wrede@triumf.ca
Subject: DSSSD figures

Dave,

Could you include this attachment and text with the minutes?  Thanks.

                  -Chris

DRAGONeers,

Ahmed and I did some tests over the weekend and it appears that DRAGON
will be able to use its fried detectors without significant degradation
in performance provided that the cooling system is installed.

Here are the spectra that I was going to show at today's meeting.

There are three spectra.  Each spectrum is of a thick, uncollimated
Am-241 alpha source placed ~20cm from the surface of the detector.

The first spectrum is using DSSSD 2069-3 (fried July '02) at 17V bias,
the maximum bias that can be applied without exceeding a leakage current
of 2.0uA.     The energy resolution is 1.1% FWHM.  This is equal to the
resolution obtained with an unfried SSSSD at 50V with the same source
and geometry.  An unfried detector will not produce a signal of this
quality at 17V.  This is be explained below.  When cooled, the leakage
current drops exponentially with temperature, so the fried detector can
be run at 50V or higher if desired.  For this reason, timing resolution
will likely not be jeopardized.

The second spectrum is taken using the fried (Oct '01) DSSSD (2069-5)
that allows the least voltage for a leakage current < 2.0uA for any of
our fried detectors;  i.e. it is fried the worst.   The spectrum was
taken at a bias of 9.5V (where I=2.0uA)  at room temperature.  There are
no counts over threshold.  This detector does not produce a good signal
at room temperature.  I  include this null spectrum for comparison with
the next one.

The third spectrum is taken with the same detector (2069-5) at a cold
plate temperature of 0degC (this corresponds to a detector temp of
~3degC), for which the leakage current is 0.88uA.  The temperature and
current can be reduced further if desired (we can go as low as
-14degC).  The spectrum has a resolution of 1.2% FWHM.  Clearly, the
detector works poorly at room temperature and like new when cooled.

We have been assuming that a detector that can't hold 50V bias is
useless.  With a moderate particle fluence, the resistivity of the bulk
silicon decreases; this is clear from the increase in leakage current in
our fried detectors.  The decrease in resistivity makes it possible to
run at lower bias voltage (the electron-hole pairs have an easier time
being swept up).  This explains the good quality spectrum obtained from
the moderately fried detector 2069-3 at the low bias of 17V.

For larger fluences, acceptor states are produced in the bulk n-type
silicon which trap electrons and a negative space charge is built up.  A
higher bias than usual is required to overcome this effect (for large
fluences, V(bias) is proportional to fluence).   This is what eventually
renders a DSSSD unusable.  None of our DSSSDs have reached the point of
destruction by fluence.

Although the initial motivation for cooling the DSSSDs was improving
their energy resolution, it should rather have been the same motivation
that most groups have for cooling their DSSSDs: cooling allows one to
use a DSSSD to detect particles even after it has detected some
particles.

                                     -Chris