Efficiency of Irradiated Double-sided Silicon Strip Detectors
Forest Atkinson, Matthias Kratzer, Teela Pulliam,
Hartmut F.-W. Sadrozinski, Riko Wichmann
Author: Riko Wichmann
This page was created September 4, 1995, last modified March 4, 1996
We investigated the change in efficiency as a function of depletion voltage for
an irradiated double sided silicon strip detector. The investigation was performed
for different fluences, since the detector was non-uniformally irradiated.
Detector
The detector was a Hamamatsu AC coupled double sided silicon strip detector of 6 cm
length. The features of the detector are the following:
- 50 µm pitch
- 12 µm implants
- 6 µm metal strips
- 26 µm p-blocking implants of n-side
The detector was connected to our binary readout system with LBIC amplifier chips and
CDP64s. The irradiation was non-uniform with a fluence between 2*10^13 and 5*10^13
protons/cm^2 depending on the beam position on the detector. Position scans show
the dependence of the efficiency on the fluence as expected from the variation of the
depletion voltage.
The irradiation was followed by a period of annealing at room and elevated (35 degree C)
temperature to raise the depletion voltage.
This simulates operation after 10 years of ATLAS operation at lowered temperature.
History of the Detector concerning its Depletion Voltage
If we follow the change in the depletion voltage at the most irradiated point, we find
the following values:
- before irradiation: 77 V across the detector
- after irradiation
- and elavated temperature annealing: 110 V
- at time of March 95 KeK beam test: 120 V
- in August 95: 190 V
The bulk of the detector was inverted after irradiation fron n-type to p-type.
The pre-rad value was obtained by measurement, while the post-rad values were predicted from
our radiation damage data and confirmed by the p-side behaviour. A direct measurement using
C-V curves was not usable due to the non-uniformity of the irradiation along the strips.
Efficiency as a Function of Depletion Voltage
The efficiency of the detector was measured for 1 fC threshold by illumination with a
106Ru beam of 8 degree divergence. The beam spot on the detector was 3mmX6mm. These results
were verified in a KeK beam test for both the pre-rad and the post-rad efficiencies before and
after annealing.
- before irradiation
- ohmic side
- is shorted out below depletion voltage resulting in a rapid loss of pulse height
with decreasing voltage
- after irradiation
- now junction side due to inversion
- shows close to sqrt(V) behaviour of pulse height
- before irradiation
- junction side
- has a close to sqrt(V) behaviour of the puls height (because of the limited
collection region below depletion)
- after irradiation
- now ohmic side
- is shorted out below depletion voltage; efficiency approximately 50% at 30 V below
depletion voltage
We measured the efficiency as a function of bias voltage up to 180 V across the detector for
thresholds of
1.0 fC,
1.2 fC and
1.5 fC. For higher thresholds we need a higher bias voltage to
reach the "saturation region" (appr. 95 %) of the efficiency on the n-side. For the p-side, the
increase in thresholds results in an even higher bias voltage to get a reasonable high
efficiency. This can also be viewed in terms of the
median pulse heigth vs bias voltage which shows, that the signal
strength on the n-side is much higher than on the p-side for bias voltages below the
depletion voltage of the detector.
At ATLAS only single sided detectors will be used so
that we won't have a voltage drop across the coupling capacitor. Therefore higher bias voltages
can be applied.
The usefulness of p-side detectors is questionable due to the low efficiency after annealing even
at high bias voltages, while the n-side efficiency is still above 95% even at half the
depletion voltage.
Conclusion
For ATLAS, we expect a depletion voltage of about 130 V after a fluence of 10^14 particles/cm^2,
since the detectors will be operated at -10 degree C. From our data, it looks very plausible that
one can operate n-side detectors safely with high efficiency after at least 3 times this
fluence ( 3*10^14 particles/cm^2 ). A fluence that high would result in a depletion voltage
of 400 V.
But the former results enable us to operate the (n-side) detectors at 200 V (in this
example half the depletion voltage) with high efficiency,
if external reasons (cooling) force us to limit the voltage.
Reference
Break out of frames to bookmark page thereafter.