Results on SGS Thompson Detectors
Liverpool CDF Group
Tests on SGS Thompson Sensors
as presented by Phil Allport at
L00 meeting, Tuesday, 11 April 2000.
We have performed a series of tests on the wafer of
11 detectors which were shipped to Liverpool.
We have compared our results, firstly to the
specifications drawn up by
Nicola and Phil, and secondly to the Hamamatsu devices.
1) Strip Failures
| | 50V | 100V | |
| SGS Thomson | 2.8% | 3.7% | (on those we can measure
~2.5 detectors) |
| | | Those failing at 100V do not subsequently hold any voltage |
| Hamamatsu | <0.5% | <0.5% | (in fact we have no reason to
disagree with the manufacturers |
| | | own results which indicate <0.1%) |
Specifications:
< 1.28% bad strips. Max bad strips: 2 (type A), 1 (type B)
> 100V coupling capacitor breakdown
Some of the detectors/regions studied show the very long time constant
behaviour reported by Ronan and Tara at FNAL so for these we cannot
measure the strip failures as at short times the currents seen are up to
mA levels (these are not included in the bad-strip estimates). This behaviour
is being studied further as there is clearly some associated `memory' with
strips which have been held at high voltage for any length of time,
subsequently showing much faster recovery from the application of high
voltage.
2) Strip Resistance
| SGS Thomson | 280 Ohms ---> 36 Ohms/cm |
| Hamamatsu | 190 Ohms ---> 24 Ohms/cm |
Specification: < 30 Ohms/cm
3) IV Curves
SGS Thomson values look fantastically low. This may be true but the
characteristics look odd and we need to check the connectivity to the
strips. After irradiation the current is dominated by the induced defects
in the substrate so starting currents are not terribly significant.
Specification: < 4uA (type A) , < 2uA (type B)
4) Resistor Values
SGS Thomson 760+/-20 kOhms, 800+/-20kOhms and 840+/-20kOhms on 3 detectors
studied. On the 4th: 800kOhms for the first and last 10% of strips but with
the central 80% showing only 40kOhms. All the above are measured at 140V.
Possibly the low values could result from a conducting inversion layer
shorting out the resistors?
Specification: 2.5 +/- 0.5 MOhms
5) Interstrip Capacitance
Genuinely lower for the SGS Thomson at the frequencies we can study
(it looks to get closer to the Hamamatsu values at higher frequency)
There is voltage dependence but I would quote 1pF/cm for the Hamamatsu
and more like 0.8pF/cm for the SGS Thomson.
(If there is any problem with connectivity to the implants these results
are maybe not going to be meaningful for the SGS Thomson.)
Note ATLAS results after irradiation showed that <100> material such as
Thomson may be using did show a lower interstrip capacitance than the
<111> used by Hamamatsu but with a very strong frequency dependence such
that the difference reduces with frequency and as far as the actual
results with fast electronics are concerned (ie at 40MHz) no difference in
noise was observable for irradiated detectors.
Specification: < 1.2pF/cm total interstrip capacitance
(The total capacitive load will also include a ~0.2pF/cm contribution to the
backplane.)
6) Depletion Voltage
We have not checked the depletion voltages. We know these are supposed
to be well over 100V for the SGS Thomson but this would not in itself a major
issue except that the capacitors are not rated to the required depletion
voltages. For Hamamatsu all the results on strip failures are at 100V
which is well above the required operating voltage of the detectors.
Specification: 60V < V_dep <100V
7) Radiation Hardness
Over 2 dozen Hamamatsu detectors to these designs (but much larger
area) have been irradiated by ATLAS in charged hadrons at the CERN and KEK
PS machines with doses of 3 10^14 p/cm^2. All devices have survived and
simply show the expected rise in depletion voltage and reverse currents.
Most of these detectors have been studied after irradiation with ATLAS
fast read-out and so the noise and any failures on the 768 strips have
been looked for. The detector satisfy the ATLAS stringent specifications
which include the requirement that the number of strips failing after
irradiation remains below 1%.
Devices have also been irradiated with neutrons but charged hadrons are
mostly what we get in the experiment and there are differences.
SGS detectors have been irradiated with neutrons.
We await the results. Irradiation with charged hadrons has not yet
been performed.
8) Quality Assurance
All detectors supplied by Hamamatsu have been put by the company
through a rigorous measurement programme and we know from ATLAS that the
Hamamatsu QA is completely reliable. Micron are also to be trusted in
terms of their measurement QA. However, the SGS Thomson are delivered
untested so I would ask what checks the company has on the quality of what
it is producing.
Apparently no SGS Thomson QA?
9) Track Record
Other companies (most notable SEIKO for BELLE) have attempted to start
up in this business and, in my experience, all of them have failed for the
first one or two iterations to get devices that work. SEIKO were dropped
by BELLE because of this. I think even for a good company there is a
learning curve and I actually think the SGS Thomson results look quite
good for a first iteration but ...
Would you want to install these in an experiment to do physics?
Back to
Liverpool CDF Group Homepage
Maintained by
mcnulty@hep.ph.liv.ac.uk