Recent Progress in Low-Temperature Silicon Detectors
M. Abreul, N. D’Ambrosioo, W. Bellg, P. Berglundj, E. Borchie, W. de Boerk, K. Boreri, M. Bruzzie,
S. Buontempoo, L. Casagrandec, S. Chapuyf, V. Cindrom, S.R.H. Devineg, B. Dezilliea, A. Dierlammk,
Z. Dimcovskif, V. Ereminp, A. Espositon, V. Granatab, E. Grigorievk, S. Grohmannc, F. Haulerk, E. Heijnec,
S. Heisingk, O. Hempeld, R. Herzogd, J. Härkönenh, S. Janosi, L. Jungermannk, I. Konorovn, Z. Lia, C. Lourençoc,
R. De Masin, D. Menichellie, M. Mikuzm, T.O. Niinikoskic, V. O’Sheag, S. Paganoo, V.G. Palmieric, S. Pauln,
K. Pretzli, K. Smithg, B. Perea Solanoc, P. Sousal, S. Pirolloe, P. Rato Mendesl, G. Ruggierog, P. Sondereggerc,
E. Tuominenh, E. Verbitskayap, C. da Viáb, S. Wattsb, E. Wobstd , M. Zavrtanikm
a
Brookhaven National Laboratory, Upton, NY 11973-5000, USA
Brunel University, Uxbridge, Middlesex UB8 3PH, UK
c
CERN, CH-1211 Geneva, Switzerland
d
ILK Dresden, Bertolt-Brecht-Allee 20, 01309 Dresden, Germany
e
Dipartimento di Energetica, Universitá di Firenze, I-50139 Firenze, Italy
f
Department de Radiologie, Universite de Geneve, CH-1211 Geneva, Switzerland
g
Department of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK
h
Helsinki Institute of Physics, P.O.Box 64, 00014 University of Helsinki, Finland
i
Laboratorium für Hochenergiephysik der Universität Bern, Sidlerstarsse 5, CH-3012 Bern, Switzerland
j
Low Temperature Laboratory, Helsinki University of Technology, FI-02150 Espoo, Finland
k
IEKP University of Karlsruhe, D-76128 Karlsruhe, Germany
l
LIP, Av. E. Garcia, P-1000 Lisbon, Portugal
m
Jozef Stefan Institute, Exp. Particle Physics Dep., PO Box 3000, 1001 Ljubljana, Slovenia
n
Physik Department E18, Technische Universität München, D-85748 Garching, Germany
o
Dipartimento di Fisica, Universitá "Federico II" and INFN, I-80125 Napoli, Italy
p
Ioffe Physico-Technical Institute, Russian Academy of Sciences, St. Petersburg 194021, Russia
b
The CERN RD39 Collaboration studies the possibility to extend the detector lifetime in a hostile radiation
environment by operating them at low temperatures. The outstanding illustration is the Lazarus effect, which
showed a broad operational temperature range around 130 K for neutron irradiated silicon detectors.
1.
INTRODUCTION
The CERN RD39 Collaboration studies the
possibility to extend the detector lifetime in a hostile
radiation environment by operating them at low
temperatures. The main goal of the collaboration is
to demonstrate full-size detector modules taking
advantage of the radiation hardness of silicon
detectors, improved by a factor of 10 or more. The
physical approach is an optimisation of the electric
field distribution in deep level rich semiconductors
that can be achieved in silicon at low temperature.
The outstanding illustration is the Lazarus effect,
which showed a broad operational temperature range
around the temperature of 130 K for neutron
irradiated silicon detectors [1].
The low temperature operation entails additional
benefits based on the increased charge carrier
mobility, on the increased thermal conductivity and
on the reduced bulk-generated current. These lead to
new design optimisations and enable to develop
segmented silicon detectors with faster signal and
higher signal-to-noise ratio. The CMOS readout
electronics will also be faster and will feature input
noise minimum around 110 K temperature.
RD39 shares resources and know-how with the
on-going CERN experiments in implementing
radiation hard cryogenic detectors and electronics.
The achievements of these projects, and the results
of the Device Physics and Basic Research projects,
were recently reviewed [2].
Important new results, to be discussed here, have
been obtained in the development of edgeless silicon
sensors [3] and in the thermal and mechanical tests
of glued silicon module structures, in the framework
of the TOTEM experiment (Total Cross Section,
Elastic Scattering and Diffraction Dissociation at the
LHC). In collaboration with the CERN experiment
NA60 (Study of Prompt Dimuon and Charm
Production with Proton and Heavy Ion Beams at
SPS) primary beam telescopes were constructed for
protons and lead ions, where the proton beam
telescope featured fast cold Deep SubMicron (DSM)
APV25 preamplifiers connected to the microstrip
sensors. The COMPASS (COmmon Muon Proton
Apparatus for Structure and Spectroscopy)
experiment will also use cold DSM readout circuits
in their cryogenic microstrip modules, where the
speed improvement was already demonstrated in the
APV25 chips. The latest activities in RD39 include
studies of particle detectors made on silicon wafers
grown by Magnetic Czochralski (MCZ) method.
2.
DEVELOPMENT
SILICON SENSORS
OF
hand scribing. The dicing was performed from either
the front p+ side or the back n+ side. The detectors
had an active area of 0.25 cm2 (5x5 mm2) before
dicing, reduced to 0.228 cm2 after dicing. Figure 1
shows photographs of a laser diced silicon
microstrip detector.
EDGELESS
In the Roman Pot detector set-up in the TOTEM
experiment at CERN [4], two halves of silicon
microstrip detectors come together so that the centre
area (a circular hole in this case) is as close as
possible to the LHC beam, and a minimum or even
zero dead area at the edge of the detector is desired.
These dead spaces are normally needed to prevent
high surface current and breakdown when biasing
above the full depletion voltage [5, 6]. It was
inconceivable until this work [3] that one can extend
the p+ implanted sensitive area all the way to the
dicing edge, practically leaving no dead space on
that edge (“edgeless” detectors, or ED).
Before cutting the p+/n/n+ silicon microstrip
detectors, the leakage current was typically of the
order of a few nA at full depletion voltage (Vfd = 85100 V), and there was no breakdown up to 500 V.
The dicing tools used in this study were a standard
laser dicing machine and a diamond point pen for
Figure 1. Photographs of a silicon microstrip
detector laser diced from the backside: a) top view,
and b) cross sectional view of the diced edge [3].
Immediately after dicing the leakage current was
increased drastically by five orders of magnitude,
with a current at full depletion voltage close to 1
mA. Due to this high leakage current, it was not
possible to yield data adequate for determining the
depletion voltage.
However, just after overnight aging at RT in air,
the leakage current improved by three orders of
magnitude for detectors diced from the backside. For
a detector diced with laser from the backside the
value of leakage current at full depletion was only
500 nA and the leakage current at 500 volts was
about 70 A. Similar results were obtained for a
detector scribed by a diamond pen from the
backside. However, there is little aging improvement
for detectors laser diced from the front side. The
improvement seen in detectors diced from the
backside is most likely due to the oxidation of the
silicon with native oxide, which passivates the cut
surface. The time constant for this aging process is
 12 hours, since longer aging time did not produce
further improvement.
Since the cutting with a scribing pen is highly
uncontrollable due to the unpredictable breaking
direction, and it can only cut straight lines, the best
choice is obviously the dicing by the laser tool from
the non-sensitive backside.
Some detectors were edge-etched 1–2 minutes in
an acid solution after dicing. The leakage current
was further decreased by more than an order of
magnitude: at full depletion (about 90 V) the leakage
current is about 0.15 A, and at V = 400 V it is
about 0.5 A.
The “edgeless” devices, diced from the backside
and after aging, also behave nicely as detectors. This
was verified by Current-Voltage measurements,
Transient Current Technique (TCT) measurements
[7] and Charge Collection Efficiency (CCE)
measurements.
3.
THERMOMECHANICAL
TESTS
OF
GLUED
MICROSTRIP
SILICON
MODULES
The TOTEM strip detector prototype module
design, shown in Figure 2, is based on single-sided
edgeless sensors mounted on a base plate together
with a pitch adapter and a hybrid readout circuit with
four APV25 chips. Between the pitch adapter and
the base plate there is a carbon fibre composite
(CFC) spacer of 0.6 mm thickness, in which a
cupronickel micropipe of 0.6 mm diameter is
integrated for providing cooling to the hybrid and
sensor of the module, while separating them
thermally.
In each Roman pot there will be two such
modules, having crossed strips for providing x- and
y-coordinates for traversing charged particles. The
APV25 chip used on the module is developed by the
CERN Microelectronics group and has 128 channels
with each channel dissipating 2.31 mW. Therefore,
the heat load per double-sided module with 10
readout chips is about 3 W. Thermal radiation occurs
between the Roman pot, which is at about 300 K,
and the detector module cooled at 130 K. According
to the calculations, one module absorbs a thermal
radiation heat load of 235 mW.
APV25
Pitch
Adapter
Hybrid
Support
Detector
Spacer
Cooling
Pipe
Figure 2. The main components of the single-sided
TOTEM module [2].
The thermal behaviour of the half-module with
vertical strips has been analysed using finite element
modelling in ANSYS. Several geometries and
material combinations were studied. The maximum
temperature difference between the fluid saturation
temperature and the readout-electronics was found
to be less than 10 K. The biggest gradient is located
within a small area around the cooling pipe. The
silicon sensor has a temperature of about 127 K and
the hybrid is only marginally warmer.
Silicon is a very good heat spreader at low
temperatures, with a thermal conductivity of about 7
W/cmK at 120 K. In order to achieve small
temperature gradients and a uniform temperature
profile in the module, we are planning to use silicon
not only for the strip detector itself, but also as a
constructional material. This also avoids the
generation of mechanical stress or deformation
during cool-down due to differences in thermal
expansion.
The thermal dilatation of standard epoxies is one
order of magnitude higher than that of silicon.
Therefore, when cooling down silicon-to-silicon
joints a lot of stress is induced. This may result in
either the joint or substrate cracking. Several models
describing matrix composite properties predict a
lower thermal dilatation coefficient for epoxies that
are charged with low thermal dilatation particles
such as fused quartz. Thus, a series of tests to study
the thermo-elastic properties of charged epoxies are
being conducted. Preliminary results prove that we
can reduce significantly the thermal dilatation of an
epoxy by filling it with an appropriate powder.
4.
CLOSED CYCLE COOLING SYSTEM
In the Roman Pot detectors of the TOTEM
experiment, cryogenic cooling is mainly needed for
the reduction of the edge surface current [8].
Considering a power dissipation of 3 W per module,
the only practical solution is to have one
autonomous cooling system for each individual
Roman pot station. With four detector modules per
station and all the thermal losses in the system the
overall cooling power amounts to 20 W at 130 K.
Our cooling system concept is shown in Figure 3.
The cooling capacity generated by a cryocooler is
distributed to the detector modules in a closed
capillary tube circuit by means of 2-phase flow of
argon. A cryogenic micropump, which is a new
prototype design, generates a suitable pressure head
to pump the fluid to the detector modules and
compensate for all the pressure losses in the circuit.
Figure 3: Cryogenic cooling system for the Roman
pot stations [9].
The novel cryogenic micropump, see Figure 4, is
based on a commercial micro annular gear pump for
room temperature applications [10]. At a flow rate
of argon of 100 mg/s or 5.2 ml/min, Pressure-head
of almost 3 bars was generated at the maximum
pump speed of 6000 rpm. This large pressure head
enables distributed cooling at a new level of
miniaturization, entailing very small thermal losses,
highly efficient heat exchangers and simple control.
The results suggest an optimum capillary diameter
of 0.3 mm. After a series of experiments and
modifications the thermal management of the
cooling circuit is now well understood.
Figure 4. Prototype cryogenic micropump [11].
5.
BEAM TELESCOPES FOR PROTONS
AND LEAD IONS (NA60)
In collaboration with the CERN experiment NA60
[12], primary beam monitoring telescopes were
constructed for protons and lead ions based on
silicon microstrip detectors operated at 130 K.
Prototype lead ion beamscope was installed in the
NA60 beam line. Inside the vacuum chamber, there
are two tracking stations, each consisting of a
printed circuit board module with two silicon
microstrip detectors mounted back to back, with the
strips perpendicular to each other. The chamber is
closed by a top flange, which contains apertures for
the detector modules, the liquid nitrogen transfer
line, and instrumentation connections. Liquid
nitrogen flows from a transport dewar vessel through
a low-loss transfer line, and feeds capillary pipes
soldered on the detector module. The temperature of
the detectors can be adjusted between 80 K and
300 K by varying the nitrogen flow in the pipes and
the power dissipated through a heater placed on the
PCB.
The lead ion telescope was exposed to the CERNSPS high intensity lead-ion (Pb) beam. The
telescope was successfully operated up to the dose
of 0.9 GGy in the centre of the beam spot.
The main difference between the proton and the
heavy ion beam telescopes is the CMOS front-end
amplifier chip placed on the proton detector module,
one per plane, close to the microstrip sensor. The
chip provides fast amplification of the current
signals generated by the protons in the microstrip
sensors. The front-end chip is manufactured in Deep
Submicron (DSM), 0.25 µm CMOS technology. It
was optimised for operation at 130 K to achieve a
fast shaping time and low input noise figure. Figure
5 shows the module design, including the front-end
CMOS chip and the cooling pipes. In the tests, the
fast cold DSM chips demonstrated cryogenic
acceleration and lowered noise at the operating
temperature of 130 K.
cooling pipe
The tests performed show that it is possible to use
a CMOS technology read-out chip at cryogenic
temperatures. In addition, the APV25 signals are
faster and higher at 130 K than in room temperature,
i.e. cryoacceleration was detected.
7.
CMOS
detector
Figure 5. Detector module for the proton beam
telescope [2].
6.
COLD
SUBMICRON
CIRCUITS FOR SILICON
(COMPASS)
READOUT
MODULES
The COMPASS experiment [13] will use
segmented double-sided silicon detectors for particle
detection. The APV25 0.25 m CMOS read-out chip
was tested at cryogenic temperatures. The chip
properties were measured at room temperature and
at 130 K. Figure 6 shows the output signal
corresponding to 1 MIP equivalent input at room
temperature (RT) and at 130 K in two chip operating
modes.
Adding oxygen into silicon material is believed to
improve the radiation hardness of silicon detectors
[14]. Silicon wafers grown by Czochralski method
intrinsically contain high level of oxygen, typically
1017 – 1018 cm-3.
We have processed, tested electrically, tested in a
particle beam, and irradiated strip detectors made on
material grown by magnetic Czochralski (MCZ)
method. The detector process is a simple four mask
level process. The measurements of electrical
performance showed leakage current of 3 uA at 900
V for the 32.5 cm detector. The depletion voltage
for the 380 um device was 420 V. Additionally, no
breakdown was detected below 1000 V.
The detection performance was tested in a muon
beam using Helsinki beam telescope [15]. The
resolution was found to be 10 µm, the efficiency
about 95 %, and the signal to noise ratio about 10.
The radiation tolerance is actually being tested by
gamma, neutron, proton irradiations. The
preliminary results show excellent radiation
hardness.
8.
Figure 6. Pulse shape corresponding to 1 MIP input
signal in two operating modes of the chip at room
temperature and at 130K [2].
SILICON DETECTORS MADE ON
MAGNETIC CZOCHRALSKI SILICON
CONCLUSIONS
Our research shows that it is possible to design
and operate radiation hard tracking detectors at
cryogenic temperatures.
Edgeless detectors were developed for the CERN
Totem experiment. Detectors diced from the nonsensitive back surface and aged one day in air
performed well as detectors. Edge etching treatment
further improved the detector characteristics.
Additionally, cryogenic micropump was succesfully
developed for the cooling system of the Totem
Silicon microstrip detector modules operating at
cryogenic temperatures.
Together with CERN Compass experiment, good
operation and cryoacceleration of the APV25 Deep
SubMicron (DSM) CMOS-chip was demonstrated at
cryogenic temperatures. With CERN NA60
experiment, beam telescopes for protons and lead
ions were succesfully operated at cryogenic
temperatures. Additionally, the proton telescope
succesfully incorperated DSM chips.
Silicon detectors made on Magnetic Czochralski
Silicon were processed, tested electrically, tested in
muon beam, and irradiated. According to the results,
the Czochralski Silicon is excellent material for
processing radiation hard silicon microstrip
detectors.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
K. Borer et al., Nucl. Instr. and Meth. A 440,
5 (2000).
K. Borer et al., RD39 Status Report,
CERN/LHCC 2002-004 (2002).
Z. Li et al., IEEE Trans. Nucl. Sci. 49, 1040
(2002).
TOTEM Technical Proposal, CERN/LHCC
99-7; LHCC/P5 (1999).
L.. Evenson, A. Hanneborg, T Happ, A.H.
Wuosmaa, and R.R. Betts, Nucl. Instr. and
Meth. A 326, 136 (1993).
Z. Li et al., Nucl. Instr. and Meth. A 409, 180
(1998).
V. Eremin, N. Strokan, E. Verbitskaya, and Z.
Li, Nucl. Instr. and Meth. A 372, 388 (1996).
S. Grohmann & al., IXth Blois Proc.
Workshop on Elastic and Diffractive
Scattering, Pruhonice near Prague, Czech
Republic, 2001, pp. 363-374.
Kryogene Kuhleinrichtung, German patent,
pending, file number 102 10 524.3.
S. Grohmann et al., ”Cooling Power
distribution from a small cryocooler”, Proc.
ICEC19, Grenoble 2002 (to be published).
Mikropumpe fur kryogene Flussigkeiten,
German patent, pending, file number
102 10 523.5
NA60 Proposal, CERN/SPSC 2000-010,
SPSC/P316, 7 March 2000.
COMPASS Proposal, CERN-SPSLC-96-19,
SPSLC-P-297 (1996).
G. Linström et al., Nucl. Instr. and Meth. A
466 (2001) 308.
C. Eklund et al., Nucl. Instr. and Meth. A 430
(1999) 321.