The DEPFET Particle Detector Systems and Their Characterization

advertisement
Czech Technical University in Prague
Faculty of Electrical Engineering
Doctoral Thesis
November 2012
Ján Scheirich
CZECH TECHNICAL UNIVERSITY IN PRAGUE
Faculty of Electrical Engineering
Department Microelectronics
THE DEPFET PARTICLE DETECTOR
SYSTEMS AND THEIR
CHARACTERIZATION
Doctoral Thesis
Ján Scheirich
Prague, (November 2012)
Ph.D. Programme: Electrical Engineering and Information Technology
Branch of study: Electronics
Supervisor: Prof. Miroslav Husák, PhD.
Supervisor-Specialist: RNDr. Peter Kodyš, PhD.*
_____
*
Institute of Particle and Nuclear Physics,
Charles University in Prague, Faculty of Mathematics and Physics
ACKNOWLEDGMENT
I would like to thank to my supervisors Miroslav Husák and Peter Kodyš for their valuable
advice and appreciated comments, as well as, Hans-Günther Moser, Rainer Richter,
Ladislav Andricek, Jelena Ninkovic and Christian Koffmane from the Max-Planck Institute
in Munich and Zdeněk Doležal from Charles University in Prague for lot of support and
help. I would also like to thank to Christian Oswald and Andreas Wassatsch for their
contribution in the development of the test system.
This research has been supported by the Czech Grant Agency under grant no.
P203/10/0777 “Development of the Pixel Semiconductor Detector DEPFET for New
Particle Experiments”, by the Grant Agency of the Czech Technical University in Prague,
grant no. SGS10/075/OHK3/1T/13 “Testing and Characterization of the Mini-matrix
DEPFET Particle Detectors”, grant no. SGS11/066/OHK3/1T/13 “Measurement of intrinsic
parameters of the DEPFET particle detector” and grant no. SGS12/073/OHK3/1T/13
“Innovation of the DEPFET pixel particle detector test system” and by the Ministry of
Education, Youth and Sports of the Czech Republic, grant no. LA10033 “Participation in
the Belle a Belle II experiments: study of B-meson decays”.
I declare that I wrote my dissertation thesis independently and exclusively with the use of
the cited sources.
Prohlašuji, že jsem svou disertační práci vypracoval samostatně a v předložené práci
důsledně citoval použitou literaturu.
In Prague, NOV 2 2012
Ján Scheirich
SUMMARY
An upgrade of the Belle KEK B-Factory particle experiment, Belle II, is being proceeded in
Japan. Collisions of asymmetric electron and positron beam are used for production of
pairs of B and anti-B mesons. Belle II experiment plans to confirm the existence of new
particles by studies of flavour physics reactions with high statistics. The new type of
semiconductor pixel detector the DEPFET (DEPleted Field Effect Transistor) proposed by
Kemmer and Lutz in 1987 is being used in the Belle II detector measuring type, track,
momentum and energy of the newly-originated particles. The DEPFET sensor has a
MOSFET (Metal Oxide Semiconductor Field Effect Transistor) integrated in each pixel and
introduces new concept of active pixels with low noise at room temperature, nondestructive repetitive readout and thinned technologies.
The DEPFET test system called Mini-matrix system was designed for testing and
characterization of small DEPFET prototypes. The system has overall noise around 20
electrons of equivalent input referred noise charge. It can drive sensors with 48 active
pixels by the reconfigurable steering sequences with high time resolution. The test system
integrates a computer controlled positioning of a laser beam focuser which can scan the
detector surface and computer controlled power supplies, so automated tests can be
performed. The system is placed in the thermally stabilized test chamber, which guarantees
precise measurements. Two generations of the DEPFET sensor PXD5 and PXD6 were
tested on the Mini-matrix test system in years 2010-2012 and the results were used for
characterization of the sensors and new detector designs.
Electrical characteristics of the DEPFET sensor, the studies of the bulk voltage impact on
the charge loss at the edges of the sensor and in-pixel studies on charge collection were
investigated using the Mini-matrix test system. Linearity of the DEPFET sensor was proved
with pulse width modulated laser and the sensor was calibrated with radioactive sources.
DEPFET gated operation was tested on the Mini-matrix system for the first time. DEPFET
sensor integrates charge continually in the standard operation. Gated mode of the DEPFET
allows making sensor insensitive for incoming radiation for defined time interval. The
charge previously stored in the internal gate is saved and integration can continue
afterwards. Such fast mechanism which can define a time window, where detector stops
integration of new charge, can be used for example to select out noisy bunches injected in
an accelerator. This mechanism fundamentally increases scope of applications of the
DEPFET sensor and has no parallel in other world pixel sensors.
The measurements of the DEPFET sensors performed on the Mini-matrix test system
leaded to unique results and deeper understanding of the DEPFET. Development of
extended version of the Mini-matrix system already started and it is expected, that the new
version will offer even more detailed measurements of the DEPFET sensors. DEPFET
Collaboration is leading to the end of the development of the active pixel detector and the
expected start date is in 2015. Development of the Mini-matrix system, test techniques and
obtained results are presented in this thesis.
ANOTACE
Probíhající inovace japonského částicového experimentu Belle v KEK B-Factory, Belle II,
bude využívat srážek asymetrických svazků elektronů a pozitronů k produkci párů B a antiB mezonů. Potvrzení existence nových částic studiem fyziky těžkých kvarků měřeními
s velkým množstvím událostí se plánuje v experimentu Belle II. V detektoru Belle II,
který měří typ, dráhy, moment hybnosti a energii nově vzniklých částic, bude využit nový
typ křemíkového pixelového detektoru DEPFET (DEPleted Field Effect Transistor),
předpovídaného v roce 1987 Kemmerem a Lutzem. DEPFET senzor, který obsahuje
MOSFET (Metal Oxide Semiconductor Field Effect Transistor) integrovaný v každém
pixelu, představuje nový koncept aktivního pixelu s nízkým šumem za pokojové teploty,
nedestruktivního opakovaného vyčítání a ztenčované technologie.
Na testování a charakterizaci prototypů DEPFET senzorů byl navržen test systém nazvaný
Mini-matrix system. Tento systém má celkový šum 20 elektronů ekvivalentního vstupního
nábojového šumu. Systém dokáže ovládat 48 aktivních pixelů senzoru pomocí
konfigurovatelných sekvencí s vysokým časovým rozlišením. V systému je integrována
počítačem řízená polohovatelná optická hlavice umožňující skenování povrchu detektoru
laserovým svazkem a počítačem řízené napájecí zdroje, pro automatická měření. Dvě
generace detektorů PXD5 a PXD6 byly testovány na Mini-matrix systému v letech 20102012 a výsledky byly použity k charakterizaci senzorů a pro nové návrhy detektorů.
Na Mini-matrix systému byly proměřeny elektrické charakteristiky DEPFET senzoru, byly
provedeny studie závislosti ztráty náboje na hranách detektoru na bulk napětí a odezvy na
sběr náboje v rámci jednotlivých pixelů. Linearita DEPFET senzoru byla potvrzena pulzním
laserem s modulací šířky pulzů a senzor byl kalibrován radioaktivními zdroji záření. Provoz
DEPFET senzoru v „hradlovém režimu“ (gated operation) byl poprvé testován na Minimatrix systému. Ve standardním provozu DEPFET senzor integruje náboj kontinuálně.
Hradlový režim provozu DEPFET senzoru umožňuje znecitlivění senzoru na příchozí
radiaci na přesně definovaný časový interval. Náboj uložený v interním hradle senzoru
zůstane zachován a integrace náboje může pokračovat po ukončení necitlivé fáze. Tento
rychlý mechanizmus, kterým lze definovat časové okno, v němž detektor neintegruje nově
generovaný náboj, může být například využit k odstínění nově injektovaných shluků částic
v urychlovači, které způsobují rušení detektoru. Hradlový režim provozu zásadním
způsobem rozšiřuje využití DEPFET senzoru a nemá obdobu u jiných světových pixleových
senzorů.
Měření DEPFET senzorů provedená na Mini-matrix systému vedla k jedinečným
výsledkům a hlubšímu porozumění DEPFET senzoru. Vývoj nové rozšířené verze Minimatrix systému byl zahájen a očekává se, že přinese ještě detailnější měření DEPFET
senzorů. Vývoj aktivního pixelového detektoru vedeného DEPFET Collaboration se blíží ke
konci a očekávaný start experimentu je v roce 2015. Vývoj Mini-matrix systému,
testovacích technik a dosažených výsledků je předmětem této práce.
Table of Contents:
1
INTRODUCTION........................................................................................................................... 10
1.1 Aims of the Thesis .............................................................................................................................................................11
2
BELLE II EXPERIMENT.................................................................................................................13
2.1 Physics Motivation ........................................................................................................................................................... 13
2.2 SuperKEKB Accelerator................................................................................................................................................... 15
2.2.1 Injection Noise ............................................................................................................................................................... 17
2.3 Belle II Detector ............................................................................................................................................................... 18
2.4 Vertex Detector ............................................................................................................................................................... 20
2.4.1 Silicone Vertex Detector ............................................................................................................................................... 20
2.4.2 Pixel Detector................................................................................................................................................................. 21
2.5 Interactions in the Silicone Detector...............................................................................................................................22
2.5.1 Particle Interactions ......................................................................................................................................................23
2.5.2 Photon Interactions .......................................................................................................................................................25
3
DEPFET SENSOR .......................................................................................................................... 28
3.1 Detection Principles .........................................................................................................................................................29
3.2 Clearing Process ............................................................................................................................................................... 31
3.3 DEPFET Readout .............................................................................................................................................................33
3.3.1 Source-follower ..............................................................................................................................................................33
3.3.2 Drain Readout................................................................................................................................................................34
3.4 Layouts .............................................................................................................................................................................35
3.4.1 PXD5 Generation ...........................................................................................................................................................35
3.4.2 PXD6 Generation...........................................................................................................................................................36
3.4.3 Noise Sources in the DEPFET Sensor...........................................................................................................................37
4
DEVELOPMENT OF THE DEPFET MINI-MATRIX TEST SYSTEM .................................................. 39
4.1 Test System Design .......................................................................................................................................................... 41
4.1.1 System Version V2 ......................................................................................................................................................... 41
4.1.2 Sensor Signal Acquisition..............................................................................................................................................44
4.1.3 Averaging .......................................................................................................................................................................46
4.1.4 Correlated Double Sampling .........................................................................................................................................47
4.1.5 Software Package .......................................................................................................................................................... 50
4.1.6 Test Environment .......................................................................................................................................................... 51
4.1.7 Test System Characterization........................................................................................................................................52
4.2 Discussion of the Test System Design .............................................................................................................................56
5
MEASUREMENTS......................................................................................................................... 60
5.1 Data Analysis Steps ......................................................................................................................................................... 60
5.2 Noise Measurements........................................................................................................................................................ 61
5.3 Radioactive Source Measurements ................................................................................................................................. 61
5.4 Laser Measurements ........................................................................................................................................................63
6
RESULTS ...................................................................................................................................... 67
6.1 Electrical Characteristics .................................................................................................................................................67
6.1.1 MOSFET Related Characteristics..................................................................................................................................67
6.1.2 Clear Related Characteristics ........................................................................................................................................ 71
6.1.3 Other Characteristics .....................................................................................................................................................75
6.1.4 Edge Effect .....................................................................................................................................................................75
6.1.5 Drift Regions .................................................................................................................................................................. 77
6.1.6 Discussion of the Electrical Characteristics..................................................................................................................79
6.2 System Noise ................................................................................................................................................................... 80
6.3 DEPFET Linearity ............................................................................................................................................................ 81
6.4 Calibration....................................................................................................................................................................... 82
6.5 DEPFET Gated Operation .............................................................................................................................................. 84
6.5.1 Charge Loss Measurement ............................................................................................................................................85
Page 6/116
The DEPFET Particle Detector Systems and Their Characterization
6.5.2 Junk Charge Selectivity .................................................................................................................................................87
6.5.3 Average Junk Charge Selection.....................................................................................................................................92
6.5.4 Discussion of the Gated Mode.......................................................................................................................................94
7
CONCLUSIONS ............................................................................................................................. 95
7.1 Author’s Contribution and Original Results ...................................................................................................................97
8
REFERENCES ............................................................................................................................... 99
Appendix A List of Candidate’s Works Relating to the Doctoral Thesis ...........................................................................103
Appendix B RÉSUMÉ..........................................................................................................................................................106
Appendix C Mini-matrix V2 Documentation ..................................................................................................................... 107
Appendix D Schematics........................................................................................................................................................ 111
Page 7/116
The DEPFET Particle Detector Systems and Their Characterization
ABBREVIATIONS
ADC
Analogue to Digital Converter
ADU
Analog-to-Digital Unit
AGND
Analog GrouND
ASIC
Application-Specific Integrated Circuit
ATLAS
A Toroidal LHC Apparatus, CERN Switzerland
CCD
Charge-Coupled Device
CDC
Charge Drift Chamber
CDS
Correlated Double Sampling
CERN
European Organization for Nuclear Research, Switzerland
CMOS
Complementary Metal–Oxide–Semiconductor
CMS
Compact Muon Solenoid, CERN Switzerland
CURO
Current Readout Circuit
DAC
Digital to Analogue Converter
DAQ
Data AcQuisition
DCD
Drain Current Digitizer
DEPFET
DEPleted Field Effect Transistor
DEPMOS
DEPleted Metal–Oxide–Semiconductor
DESY
Deutsches
Elektronen-SYnchrotron,
experimental
laboratory
in
Hamburg, Germany
DHP
Data Handling Processor
DSSD
Double-side Silicone Strip Detectors
ELC
ELectromagnetic Calorimeter
FPGA
Field-Programmable Gate Array
ILC
International Linear Collider
IR
Infra Red
KEK
High Energy Accelerator Research Organization, experimental laboratory
in Tsukuba, Japan
LEP
Large Electron Positron Collider, CERN Switzerland
LHC
Large Hadron Collider, CERN Switzerland
Page 8/116
The DEPFET Particle Detector Systems and Their Characterization
MIP
Minimum Ionizing Particle
MOS
Metal Oxide Semiconductor
MOSFET
Metal Oxide Semiconductor Field Effect Transistor
MPI
Max-Planck Institute, Munich, Germany
PCB
Printed Circuit Board
PCI
Peripheral Component Interconnect
PID
Particle Identification Detector / Proportional–Integral–Derivative
PS
Proton Synchrotron, CERN Switzerland
PXD
PiXel Detector
RMS
Root Mean Square
SDC
Silicone Drift Chamber
SPI
Serial Peripheral Interface Bus
SPS
Super Proton Synchrotron
SVD
Silicone Vertex Detector
TIA
TransImpedance Amplifier
USB
Universal Serial Bus
Page 9/116
The DEPFET Particle Detector Systems and Their Characterization
1
INTRODUCTION
Search for origins of matter and fundamental interactions is as old as the first
philosophy and natural sciences. Development of the Standard Model of elementary
particles and fundamental interactions was a great achievement of the last century.
Beyond the Standard Model still remain many open questions and physicists search for
answers as what is the nature of dark matter, unification of forces, how do particles
acquire mass and many others. In the last few decades, many groundbreaking physics
discoveries were done in high energy physic laboratories as CERN in Switzerland,
Fermilab in the USA, KEK in Japan and others. New accelerators and particle physic
experiments are planed in the future.
Particle physicist use linear or cyclic accelerators to accelerate charged particles. The
particles are collided at an interaction point surrounded by various types of detectors
measuring type, track, momentum and energy of the newly-originated particles. The
detector closest to the interaction point is called vertex detector and usually built of
semiconductor strip and/or pixel sensors. In 1987 Kemmer and Lutz proposed a new
type of pixel detectors: the DEPFET (DEPleted Field Effect Transistor) [ 1 ]. Today,
DEPFET based vertex pixel detectors are developed by an international collaboration
for the Belle II experiment in Japan [ 2 ] and the future International Linear Collider
[ 3 ]. The DEPFET Collaboration [ 4 ] unites universities and institutes from Germany,
Spain, the Czech Republic, Poland, Japan and China. The achievements of the particle
physics would not be possible without a significant engineering research and state-ofthe-art technologies. Development of the DEPFET pixel detector does not include only
sensor itself, but also development of steering and readout electronics, development of
test systems and testing, characterization and optimization of the sensor.
Various generations and different layouts of the DEPFET sensor were designed during
its evolution. New structures were integrated into the sensor pixels to improve the
charge collection over enlarged pixels and arrangement of the pixels was changed. All
changes of the senor needed to be tested. Also new operation of the DEPFET senor
Page 10/116
The DEPFET Particle Detector Systems and Their Characterization
called gated mode was developed to prevent sensor saturation during the injection of
new particles into the accelerator. Such new operation also needed to be tested,
optimized and verified.
Various test systems were designed [ 5 ], [ 6 ], usually based on ASICs (Application
Specific Integrated Circuits) primarily designed for the final sensor application. A new
system has been developed for testing and characterizing small samples of the DEPFET
detector, based on the components of the shelf. The system is an alternative to the
steering and readout ASICs [ 5 ], [ 6 ] and [ 7 ]. The system enables precise and lownoise charge measurements and a flexible high resolution configuration of control
signals and was used for DEPFET sensor testing.
In this thesis, DEPFET pixel detector is presented in the context of physics background
and Belle II experiment. Development of a specific test system is shown and
characterization of the two generations of the DEPFET sensor is presented. The
DEPFET gated mode operation, which is crucial for DEPFET operation in the
accelerator environment and was tested and quantified on the designed test system for
the first time, is also presented.
1.1
Aims of the Thesis
The two major goals of this project were:
•
Development of an alternative precise low-noise test system for the DEPFET
prototypes
•
Testing and characterization of the DEPFET prototypes on the previously
designed test system
Need of a new test system was urged since 2007 in the DEPFET Collaboration by
necessity of the ASIC independent system which could handle at least small area of the
DEPFET sensor. The primary requirements of the system were following:
Page 11/116
The DEPFET Particle Detector Systems and Their Characterization
•
Easy replacement of the tested devices
•
Design independent of ASICs
•
At least 8 drain channels
•
6 gate and clear channels
•
Noise lower than 20 electrons of equivalent noise charge (ENC)
•
Universal sequencer and high time resolution configuration of the control
sequences (<20 ns)
•
Broad amplitude levels of the gate and clear pulses
•
Fast design
The requirement of the characterization of the test system itself and its validation is
obvious. The basic required measurements on the test system were following:
•
Test of basic functionality
•
Measurement of electrical characteristics (characterizing) and search for optimal
operation point
•
Calibration
Other goals have showed up later on, related to the specific DEPFET prototype or
accelerator issues. These were:
•
Optimization of the drift voltage and in-pixel studies
•
Verification of the DEPFET gated mode operation and its quantification
Page 12/116
The DEPFET Particle Detector Systems and Their Characterization
2
BELLE II EXPERIMENT
The SuperKEKB-Belle II experiment [ 8 ] is an upgrade of the KEK B-Factory
experiment [ 9 ], at KEK laboratory in Japan . The KEK B-Factory experiment has been
exploring the problems of matter-antimatter asymmetry by colliding beams of highenergy electrons and positrons. The KEK B-Factory experiment has a double-ring
collider, SuperKEKB [ 10 ]. The electron and positron beams are accelerated to near the
speed of light in the KEKB's 3-kilometer long rings and collided at the center of the
Belle II detector, where produce pairs of B and anti-B mesons.
2.1
Physics Motivation
The theory, which describes the system of known particles and interactions, known as
Standard Model [ 11 ], [ 12 ] and [ 13 ], was developed in the second half of 20th century
by many physicists. The current formulation was finalized in 1970’s and later, it was
confirmed by discovery of the bottom quark in 1977, the top quark in 1995, the tau
neutrino in 2000 and the Higgs-like boson in 2012. Although the confirmation that
properties of the last discovered particle match that of the Standard Model Higgs is still
needed, the existence of such a particle strongly supports the validity of the Standard
Model. Summary of the Standard Model particles is shown in Figure 1. First generation
of elementary particles (u, d and e) constructs atoms and νe neutrino comes out from
the beta decay. The second and third generation is identical to the fist one except for
different quantum numbers (flavours) and mass. Every elementary particle has an
antiparticle, which has inverted quantum numbers.
The fundamental interactions are described in the Standard Model as exchange of
gauge bosons:
•
Photons – electromagnetic interaction
Page 13/116
The DEPFET Particle Detector Systems and Their Characterization
•
W and Z bosons – weak interactions
•
Gluons – strong interactions
Discovery of the Higgs-like particle at ATLAS and CMS experiments was pronounced at
CERN press conference on 4 July 2012. “We have reached a milestone in our
understanding of nature,” said CERN Director General Rolf Heuer. “The discovery of a
particle consistent with the Higgs boson opens the way to more detailed studies,
requiring larger statistics, which will pin down the new particle’s properties, and is
likely to shed light on other mysteries of our universe.” Beyond the Standard Model
still remain many open questions as unification of forces, imbalance between matter
and antimatter in the Universe or the neutrino masses.
Figure 1 – Elementary Particles of the Standard Model
Belle II experiment [ 2 ] plans to confirm the existence of new particles by studies of
flavour physics reactions with high statistics. In contrast to Large Hadron Collider
(LHC) [ 14 ] accelerator at CERN, where protons are accelerated at high energies (up to
Page 14/116
The DEPFET Particle Detector Systems and Their Characterization
2 x 7 TeV) and new particles are studied directly, SuperKEKB is aiming at high
statistics allowing indirect studies of new particles. High precision of measurements at
Belle II will improve the first results of ATLAS [ 15 ]and CMS [ 16 ] experiments at
CERN.
2.2
SuperKEKB Accelerator
Energies of particles, which naturally come out of alpha and beta decays, are too low to
be used in particle experiments. Ernest Rutherford urged the need of first proton
accelerator with his research and Ernest Lawrence invented a cyclotron accelerator in
1929, based on the studies of Rolf Widerøe.
Today exist two possible approaches how to construct large accelerators. Circular
synchrotron topology is used at CERN facilities, as Large Hadron Collider (LHC),
Proton Synchrotron (PS) and Super Proton Synchrotron (SPS). Previously, Large
Electron Positron Collider (LEP) at CERN was installed. In the USA, the Fermilab
proton anti-proton collider TeVatron was constructed. Smaller accelerators are in
Germany (DESY), Japan (KEK) and other countries. In these colliders, particles are
accelerated in a ring until they reach a designated energy and they are brought into
collision than.
The synchrotron radiation emitted by charged particle flying on the circular trajectory
is proportional to 1/m4. The energy loss ∆E due to synchrotron radiation for one
circular revolution of a highly relativistic particle with rest mass m0 and energy E is
given by[ 5 ]:
∆E =
(
e2
3ε 0 m0 c 2
)
4
E4
,
r
Equation 1
Page 15/116
The DEPFET Particle Detector Systems and Their Characterization
where e is the elementary charge, ε0 is the dielectric coefficient, c is the speed of light
and r is the radius of the circular path. The synchrotron radiation is a limiting factor
for reaching high energy collisions in circular accelerators. Different concept proposed
for high energy acceleration is a linear collider. International Linear Collider (ILC) is
planned to be built in linear topology. In this case two particles start at the different
places and they are accelerated in a straight line before they collide. There are no
energy losses due to the synchrotron radiation, but the particle has to be accelerated to
its desired energy in a single passage through the accelerator.
Figure 2 – Schematic Layout of SuperKEKB Accelerator [ 10 ]
The SuperKEKB is a 3-kilometer long cyclic synchrotron accelerator. Figure 2 shows
the scheme of the accelerator. It is an upgrade of the previous KEKB collider. Most of
the commitments will be re-used. However, many components need to be redesigned
and newly developed. Need of high rate of events N will be accomplished by high
luminosity L of the accelerator:
Page 16/116
The DEPFET Particle Detector Systems and Their Characterization
N = Lσ
,
Equation 2
where σ stands for cross section of the particle interaction. High luminosity of the
machine is achieved by “nano-beam” where particles are compressed and collided with
high precision at the interaction point. The target luminosity (1~5.1035 cm-2s-1) of the
SuperKEKB is almost ten times higher than luminosity of the previous KEKB.
Figure 3 – Scheme of Injection at SuperKEKB, data [ 2 ]
2.2.1
Injection Noise
SuperKEKB accelerator is using continuous injection of new bunches of particles, to
keep luminosity at the constant level. Each newly injected bunch moves first 4 ms on
unstable trajectories emitting “noise” radiation which might interfere with the pixel
detector. Two bunches are injected 100 ns apart every 20 ms and after 4 ms of cooling
there are 16 ms of clean operation. One revolution of the 3 km long accelerator takes
10 µs. Exact radiation of the noisy bunches is not known yet, but rough estimation was
done by C Kiesling at 11th Belle II General Meeting and charge 6.105 should be
generated in one detector’s pixel during passage of two noisy bunches. Since the
integration time of the DEPFET detector is 20 µs, it means, that each readout frame of
the DEPFET integrates 4 passages of the noise bunches. DEPFET would be completely
saturated by such amount of collected charge and it would lead to 20 % data loss. To
prevent such data loss due to the injection noise issue the DEPFET gated mode
Page 17/116
The DEPFET Particle Detector Systems and Their Characterization
operation were developed and tested. Tests first tests of DEPFET gated mode operation
carried on the Mini-matrix test system are reported in Chapter 6.5 of this thesis.
Figure 4 – Belle II Detector [ 8 ]
2.3
Belle II Detector
Belle II detector is an upgrade of the previous Belle detector constructed for high
luminosity beam. Figure 4 and Figure 5 show schematic drawing of the Belle II
detector. Newly generated particles created in collisions of electrons and positrons fly
through different detector layers. The detector closest to the interaction point is called
the vertex detector. It is composed of two layers of the DEPFET pixel detector (PXD)
and four layers of double side silicone strip detectors (SVD). Vertex detector is used for
precise determination of the decay vertex of the particles produced in collision. Tracks
of charged particles are curved in the magnetic field, when they fly through the central
Page 18/116
The DEPFET Particle Detector Systems and Their Characterization
drift chamber (CDC). Momentum of these particles can be measured in CDC based on
the diameter of particles tracks. Particle identification detector (PID), composed of
endcap and barrel, is used to distinguish among different particle types. Layer of
electromagnetic calorimeters (ELC) follows after layer of PID. ELC is used for
measurement of particle energy. The last layer is KL meson and muon detector.
Figure 5 – Belle II Detector Close Up [ 8 ]
The selected components have to sustain an environment with considerably higher
radiation background levels than in the previous Belle experiment. In compare to
Belle, the Belle II expects increase of background by factor of twenty and the physics
event rate by factor of fifty [ 2 ]. The major changes of Belle II are:
•
The DEPFET Pixel detector (PXD) will be added
•
The silicon strip detectors (SVD) will be extended and equipped with new
readout chips
•
The central drift chamber will have smaller cells
•
The completely new particle identification detector will be installed
•
New electronics of electromagnetic calorimeters will be used and endcap
scintillators will be replaced
Page 19/116
The DEPFET Particle Detector Systems and Their Characterization
•
KL and muon detector will be upgraded
•
New data acquisition system will be developed and installed
2.4
Vertex Detector
The vertex detector plays major role in the measurement of B meson decays. The decay
time of B meson by weak interaction is extremely short and not possible to measure
directly. Hence, the indirect measurement will be carried out by measurement of the
decay vertex. The asymmetric beam energies causes shift in the angle of the vertex in
the direction of the more energetic electron beam.
2.4.1
Silicone Vertex Detector
Silicone vertex detector (SVD) [ 17 ] is composed of four layers of double-side silicone
strip detectors (DSSD), as it was done also in the previous Belle detector. The strip
detectors are shifted in the direction of electron beam, for better coverage of the
boosted vertex, as illustrated in Figure 5. The strip detector consists of finger detection
structures (strips), which can detect particles in 1D space dimension. The double-side
strip detectors have strips at the both sides and one is 90 degree rotated with respect to
each other. Such layout creates 2D detection device. SVD has very fast readout with
temporal granularity in order of 20 ns, but an ambiguity of multiple particle hits over
detector area will occur due to the high event rate. Two layers of innermost pixel
detector (PXD) do not suffer with spatial ambiguities, because of individual pixel
readout, but the integration time 20 µs is relatively long in compare with the event rate.
So combination of both data (SVD and PXD) is essential.
Page 20/116
The DEPFET Particle Detector Systems and Their Characterization
2.4.2 Pixel Detector
The pixel detector makes two innermost layers of the Belle II detector, the first layer
with radius 14 mm and the second with radius 22 mm. The inner layer consists of 8
modules and the outer layer of 12 modules. Each module is made of two half-modules,
which are glued in the middle. Figure 6 illustrates the arrangement of the pixel
detector. The modules are mounted on an integrated support and cooling structure by
screws. The support on the backward side can slide on the beam pipe in order to
compensate for thermal expansion of the beam pipe and the beam pipe supports.
Figure 6 – Pixel Detector [ 17 ]
The DEPFET sensor is used for the pixel detector. Very thin sensor can be
manufactured due to DEPFETs internal amplification. DEPFET module (see Figure 7)
has 75 µm thickness in the active area. (50 µm thickness of the DEPFET sensor was
achieved during the prototyping period.) The pixels itself are 50 x 55 µm and
Page 21/116
The DEPFET Particle Detector Systems and Their Characterization
50 x 60 µm large in the inner layer and 50 x 70 µm and 50 x 85 µm large in the outer
layer. Two regions with different pixel pitch are implemented to improve resolution.
The support structure around the active DEPFET sensor (balcony) is made of 450 µm
silicone. On the balcony, steering switcher chips drain current digitizer chips (DCD)
and data handling processor chips (DHP) are located. Electrical connections are
realized via a flex capton cable.
Figure 7 – DEPFET Module Structure [ 17 ]
Three different ASICs [ 18 ], [ 19 ] are mounted on the pixel detector module. Switcher
circuits are responsible for switching correct voltage sequences to the DEPFET external
gate and clear contacts. The switchers control row-wise readout of the DEPFET matrix
in the “rolling shutter mode”. Column (drain) lines of the matrix are connected to the
DCD chips. These chips process and digitize signals from the selected part of the
sensor. The raw digitized data are than processed and compressed by the DHP chips.
The HDP also provides fast serialization. All the chips are mounted directly on the
detector module leading to the fully silicone module.
2.5
Interactions in the Silicone Detector
Various materials can be used for construction of particle detectors. Silicone has several
advantages, which makes it suitable for use in high energy particle physics. Average
energy needed for creation of an electron-hole pair is only 3.65 eV. Relatively high
density of 2.33 g/cm3 causes impinging particle lose enough energy for detection even
Page 22/116
The DEPFET Particle Detector Systems and Their Characterization
in a very thin layer. High mobility of electrons and holes in the silicone makes fast
response of silicone detectors possible. Very advantageous is also fact that state-of-theart technologies exists for manufacturing silicone chips. Compatibility with front-end
electronics is than granted and readout electronics can be partially (or fully) integrated
into the detector, as it is done in case of the DEPFET detector.
2.5.1
Particle Interactions
Particle detection with semiconductor detectors is based on interaction of particles with
matter. Generally, only charged particles and photons can be primarily detected with
silicone detectors. Free charge carriers are generated (electrons and holes) which can
be converted to an electrical signal. There are several mechanisms of interaction with
matter as inelastic collisions with electrons, elastic scattering of nuclei, bremsstrahlung
(braking radiation) and Čerenkov radiation.
Heavy charged particles usually interact with the electrons of atoms of silicone via
soft inelastic collisions. In these soft collisions only small part of energy which carries
the particle is transferred and an atom of silicone is excited. In hard collisions, which
are less probable, the electron of the silicone atom is recoiled. Such interaction is
known as delta electron generation. Rarely, the energy of the delta electron is high
enough to cause secondary ionization. During the elastic scattering of nuclei only a
small amount of energy is transferred to nuclei due to the high mass difference and
such interactions plays a minor role. Energy loss of a heavy charged particle in matter is
described by the Bethe-Bloch formula:
−
dE
ρZ z 2  1  2meγ 2 v 2Wmax
= 4πN a re2 me c 2
 ln
dx
A β 2  2 
I2
 2 δ
 - β −  ,
2

Equation 3
where I is the excitation potential, Z, A, ρ are parameters of absorbing material, z is charge of
the incident particle, β is velocity of the incident particle, γ is Lorenz factor, Wmax is maximum
Page 23/116
The DEPFET Particle Detector Systems and Their Characterization
kinetic energy that can be transferred in a single collision and δ is density correction. Na, Re and
me is Avogardo’s number, classical electron radius and electron mass. In Figure 8 is depicted
application of Bethe-Bloch formula to silicone for different particles. One can see, that all
particles have approximately same minimum energy loss (at βγ = 3). For higher kinetic
energies the energy loss is growing slowly and particles can be approximated as a minimum
ionizing particle (MIP).
Figure 8 – Bethe-Bloch Formula in Silicone [ 20 ]
Electrons and positrons are much lighter and deflection of these particles has to be
taken into account by modification of the Bethe-Bloch formula. At energies higher than
0.1 GeV the bremsstrahlung mechanism of energy loss is dominant. Energy loss can be
expressed than as:
−
dE E ( x )
=
→ E ( x ) = E0 e − x X 0 ,
dx
X0
Equation 4
where X0 is a radiation length after which particle looses in matter 1/e of its initial
energy. Energy loss of a particle in matter is not exact value. There are fluctuations of
the energy loss and calculations given above are mean values of energy loss. In thick
materials the energy loss distribution is Gaussian. In a thin absorption layer, there is
Page 24/116
The DEPFET Particle Detector Systems and Their Characterization
smaller probability of collisions for particles with higher energies and energy
distribution is shifted to the lower energies as illustrated in Figure 9.
Figure 9 – Example of a Landau Distribution [ 20 ]
2.5.2 Photon Interactions
The mechanism of photon-matter interaction in silicone depends on energy of
incoming photon. Low-energy photons with energies lower than 100 keV mainly
interacts via photo effect. Photon is either completely absorbed and energy of
absorbed photon is transferred to the outermost electron of the silicone or nothing
happens and the absorbed energy is reemitted. If energy of the photon is higher than
the electron binding energy, free electron is emitted. The minimum energy necessary to
remove an electron from the valence band, called the work function Φ, is defined as:
Φ = hf 0
Equation 5
where h is a Planck constant and f0 is the minimum frequency of absorbed photon
needed for electron emission as illustrated in Figure 10.
Page 25/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 10 – Band Diagram of Silicone (Vacuum Energy EVAC, Bottom of Conduction Band
EC, Fermi Energy Level EF, Top of Valence Band EV, Electron Affinity Eea, Work Function
Φ, Band Gap Energy Eg)
When energy of the absorbed photon is high enough to excite electron from the valence
to conduction band, but not enough to leave the silicone atom, the pair electron-hole is
generated. It leads to increase of the conductivity of the silicone and to this effect is
referred as photo conductivity. The absorption length λ, which defines a distance in
material where the probability of impinging photon is not absorbed has decreased to
1/e,
P(x ) = e − x λ
Equation 6
is dependent on wavelength of the photon. Figure 11 illustrates the absorption length of
light in silicone as a function of the wavelength.
For higher photon energies, absorbed energy is higher, than electron affinity Eea and
electron is ionised (photo ionisation). The absorption length increases almost
exponentially and is much longer than a typical thickness of a silicone detector.
Efficiency of detection high energy photons drops down rapidly.
Compton scattering is a dominant effect for energies in range 0.1 – 5 MeV. In this
interaction only part of the photon energy is transferred to an electron and the photon
with lower energy escapes the silicone. In this case, only fraction of the photon energy
can be measured.
Page 26/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 11 – Light Absorption in Silicone [ 21 ]
Page 27/116
The DEPFET Particle Detector Systems and Their Characterization
3
DEPFET SENSOR
Introduction of semiconductor particle detectors to particle physics experiments causes
rapid development of new detector concepts. In 1987 Kemmer and Lutz proposed a
new type of DEPFET pixel detector [ 1 ]. It was based on double side depleted silicone
drift chambers (SCD). Kemmer and Lutz proposed an integration of MOSFETs (Metal
Oxide Semiconductor Field Effect Transistors) along the SDC surface which creates
one- or two-dimensional pixel array with non-destructive repeated readout and
amplification of the signal. They named it DEPMOS. Today, semi-monolithic DEPFET
based vertex pixel detectors derived from DEPMOS are developed by the DEPFET
collaboration for the BELLE II experiment in Japan and the future International Linear
Collider.
Figure 12 - Principle of the Sideward Depletion [ 5 ]
Page 28/116
The DEPFET Particle Detector Systems and Their Characterization
3.1
Detection Principles
Integration of MOSFET directly to each detector’s provides a first charge amplification
stage. Detector itself consists of a high-resistivity depleted n-substrate and two pregions, creating a pnp-sandwich structure (p frontside-implantation, n-substrate, pbackside). The n-substrate is sideward depleted.
The principles of sideward depletion are shown in Figure 12 and discussed in [ 5 ] and
[ 22 ]. The n-substrate (bulk) is depleted from front and back by applying negative
voltages to both p-implantations with respect to the bulk. The electron potential
minimum is in a plane parallel to the front surface. An one-dimension Poisson equation
with boundary conditions ϕ(0) = Vu and ϕ(d) = Vd describes potential distribution in xaxis:
ϕ (z ) =
eN D
z
z (d − z ) + (Vd − Vu ) + Vu
2ε s
d
Equation 7
where z is the depth in the detector substrate, e is the elementary charge, ND is the
doping concentration of the substrate, d is the total wafer thickness, εs is the dielectric
constant of the semiconductor and Vd, Vu are the voltages applied to the back and the
front side. The minimum of the potential is in the depth zmin given by:
z min =
εs
d
+
(Vd − Vu )
2 qN D d
Equation 8
When Vu = Vd , the potential minimum will be in the middle of the sensor. Asymmetric
voltages are applied to the DEPFET pixel to shift electron potential minimum close to
the front surface, where is the MOSFET. Additional n-implants hinder electron lateral
diffusion and electrons are concentrated in a small region under the MOSFET channel.
This region is called an internal gate. A principal illustration of a DEPFET pixel is
shown in Figure 13.
Page 29/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 13 - The DEPFET Operational Principles [ 17 ]
When an impinging radiation generates electron-hole pairs in the depleted n-substrate
(bulk), the holes drift to the backside contact, but the electrons are trapped in the
internal gate. The internal gate is located directly under the MOSFET channel, under
the external gate contact, so the stored charge in it affects MOSFET channel. Such
layout allows charge non-destructible reading and the reading process can be repeated
many times. For a fixed drain to source voltage VDS and a constant external gate voltage
VGS the drain current ID is proportional to the stored charge in the internal gate. The
amplification gq is given by the change of the transistor current δID due to the collected
charge δQ:
gq =
δI D
δQ
.
Equation 9
VGS ,VDS
Internal amplification can be expressed as:
Page 30/116
The DEPFET Particle Detector Systems and Their Characterization
gq = −
µh
2
L
sat
VDS
,
Equation 10
sat
where µh is the hole mobility, L is the gate length and VDS
is the transistor saturation
voltage. According to Equation 10, gq can be maximized choosing the small channel
length L.
The amplification 300 – 600 pA/e- is obtained for mini-matrices with
channel effective length of 4 µm.
3.2
Clearing Process
When new charge collection is needed, it’s necessary to empty the internal gate. For
clearing out the internal gate, there is a clear contact next to the MOSFET transistor. In
Figure 13 is SHOWN a Cleargate cross-section. Detail description of the clearing
process can be found in[ 23 ], [ 24 ] and [ 25 ]. The electrons are extracted from the
internal gate by applying high positive voltage to the Clear contact. It causes the
electrons drift to the Clear contact, where they are taken away. To prevent losses during
charge accumulation, the n+-region under the Clear contact is surrounded by p-well.
The n+-region is providing an ohmic contact to the Clear electrode and with the p-well a
reverse biased PN junction that represents a potential barrier for the electrons in the
internal gate. When the voltage applied to the Clear electrode is high enough, the
depleted region in the p-well overcome through the p-well and touches the p-well
boundary (punch-through effect [ 26 ]). In this moment there is no barrier for the
electrons in the internal gate and they are extracted.
In order to control the potential barrier between the internal gate and the Clear contact
an additional MOS structure Cleargate is added. If the Cleargate is on a positive
potential during the clear process, it helps forming an n-channel in the p-well. But
whereas the n-channel is situated at the surface, the punch-through effect is also
effective in the depths of the internal gate.
Page 31/116
The DEPFET Particle Detector Systems and Their Characterization
Several different sensor layouts make possible different Cleargate operations. It is
possible to clock the Clear and the Cleargate as shown in Figure 14. This setting needs
an extra control line for the Cleargate for each row. It’s more efficient to find a static
Cleargate operation potential. All Cleargates have than a common contact (common
cleargate), where the Cleargate voltage is kept constant during whole operation and the
clear process is controlled only by voltage at the Clear contact. Another layout
integrates a capacitor between Clear and Cleargate regions (capacitive coupled
Cleargate) and couples clear pulse to the Cleargate contact. Such layout benefits of
advantages of clocked Cleargate, but does not need an extra control lines.
Figure 14 - Readout Scheme with the Cleargate Clocking
As described in [ 25 ], the clearing process can be improved by an additional unmasked
high energy n+-implantation in the depth of 1.2 µm under the surface. Such
implantation shifts the flow of electrons during the clear process deeper to the
substrate and the punch-through of the clear is easier, because of lower capacitive
coupling between the p-well and the Cleargate. It decreases the Cleargate-voltage swing
and it allows static Cleargate-voltage operation. A negative effect of the deep-energy n+implantation is a shift of the internal gate deeper underneath the surface and
consequently reduction of the internal gain gq.
Page 32/116
The DEPFET Particle Detector Systems and Their Characterization
3.3
DEPFET Readout
3.3.1
Source-follower
The DEPFET detector makes possible two ways of readout as shown in Figure 15. The
first is a voltage based readout using a source-follower configuration. The DEPFET
transistor is biased by a constant current source and output signal is taken as a source
voltage swing. The source voltage swing ∆U is caused by charge accumulated in the
internal gate Qin. The gain G of the source-follower stage can be approximately
expressed as [ 22 ]:
G=
∆U
1
,
=
∆Qin CGD
Equation 11
where CGD is the gate-drain capacitance of the DEPFET detector. The gain is
independent of the gate-to-source capacity CGS, because VGS is constant. The setting
time τ in with the output signal reaches 63% of its maximal value is approximately
given by [ 22 ]:
 C
C L 1 + GS
CGD
τ≈ 
gm

 + CGS

,
Equation 12
where CL is a load capacity of the output node and gm is the transconductance of the
DEPFET transistor. Because CL is the capacity of the whole matrix column
(approximately 10 pF), it dominates in the numerator of Equation 12. The
transconductance gm of the transistor is limited be the gate length L technological limit
2 µm. Consequently the setting time τ is about of ones of µs and unacceptably long for
the Belle II and ILC operation.
Page 33/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 15 - DEPFET Readout (Left: Source follower; Right: Drain current readout) [ 5 ]
3.3.2 Drain Readout
In case of the drain current readout the drain-source voltage is kept constant. The
change of output current dI is given by change of accumulated charge dQin and
amplification of the internal gate gq:
dI D = g q dQin
Equation 13
In this case the setting time τ is independent of the DEPFET transistor transconduction
gm and is given by:
τ = C L Rin
Equation 14
where Rin is the input resistance of the readout electronics. In this configuration it is
possible to decrease setting time up to τ = 1 ns and it’s acceptable for the Belle II and
ILC operation.
Page 34/116
The DEPFET Particle Detector Systems and Their Characterization
3.4
Layouts
The DEPFET sensors were evolving during its development and different layout and
detector thicknesses were manufactured. These generations of the sensor are called
PXDx, as pixel detector. Not all the DEPFET generations were fabricated. Up to now
PXD4, PXD5 and PXD6 were produced and PXD9 is in preparation as a final design for
the Belle II detector. In this thesis, only PXD5 and PXD6 DEPFET generations are
described and tested.
3.4.1
PXD5 Generation
The first device tested on the Mini-matrix system (described in the following chapter)
was a detector of a PXD5 generation. It is a 450 µm thick matrix with a common clear
gate (CCG) and 20 x 20 µm large pixels. Figure 16, left, shows a layout of a DEPFET
PXD5 generation mini-matrix. PXD5 generation detector prototype was primary
designed for ILC [ 27 ], but Belle II requirements are quite similar to the ILC. Layout of
the sensor is rectangular with double pixel layout and 2-fold readout. Physical
connection of the DEPFET pixels in 2-fold and 4-fold readout is shown in Figure 17.
Two rows of pixels are activated by one gate line in 2-fold readout. Each column needs
than two drain lines to be readout. Detector is also cleared by two rows at once. The 4fold readout, which is implemented in PXD6 generation, is read out by four rows at
once. It means that one gate line activates four rows and four drain lines are needed for
one column. The parallelization in layout is implemented to achieve more relaxed
timing of MOSFETs in DEPFET pixels. The Mini-matrix system V2 has fixed number
of eight drain channels. It means that in 2-fold configuration, system can readout four
physical columns of the detector. In 4-fold configuration, only two columns can be
readout.
Page 35/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 16 – Left: PXD5 Sensor Layout; Right Photo of the Mini-matrix Sensor
Figure 17 – Physical Connection of 2 x 4 Pixels LEFT: 2-fold Readout, RIGHT: 4-fold
Readout
3.4.2 PXD6 Generation
The PXD6 generation is a DEPFET prototype designed to fulfil Bell II requirements.
The pixels for Belle II are enlarged in contrast with PXD5 which has the smallest
Page 36/116
The DEPFET Particle Detector Systems and Their Characterization
possible design. PXD6 production line contains several layout designs and thicknesses.
At the Mini-matrix system, detectors with layout shown in Figure 18 (75 x 50 µm) and
thicknesses 450 µm and 50 µm were tested. The size of the MOSFETs in the pixels stays
the same as in PXD5. Size is given by the gate length, which has to stay relatively short
to keep high gain of the MOSFETs. The active area is surrounded by empty space,
where the intensity of lateral electrical field is low. It would lead to slow electron drift in
this regions and charge lost. p+-implantations were added to fill the empty space and
they are connected to the drift contacts where drift voltage is applied to shape electrical
field inside the pixels.
Figure 18 – Left: PXD6 Sensor Layout; Right Photo of the Mini-matrix Sensor
3.4.3 Noise Sources in the DEPFET Sensor
Several noise sources can be identified in the electronic devices separated by the origin
of the noise. Electrical transport in the semiconductor devices is arranged by electrons
and holes. These carriers carry a discrete quantum of charge so current flow is not
absolutely continuous, but small fluctuation of transported charge. It is referred to the
noise caused by random fluctuations of an electron stream as a shot noise. The
spectral power density of the shot noise current is given by:
Page 37/116
The DEPFET Particle Detector Systems and Their Characterization
S ( f ) = 2e I ,
Equation 15
where e is the elementary charge of an electron, I is the average current flow and f is
the frequency. The spectrum of the shot noise is white, what means that the power
spectral density does not change with frequency.
Thermal motion of electrons in semiconductor material is a source of thermal noise.
The noise is proportional to the absolute temperature T of the semiconductor. On the
ideal resistor with resistance R the power spectral density is given by:
S ( f ) = 4kTR ,
Equation 16
where k is the Boltzmann constant. The noise spectrum of the thermal noise is also
white.
Another noise source is the 1/f noise or flicker noise. 1/f noise is present in all active
devices so as well in the DEPFET sensor. It is caused mainly by the trapping centres in
the MOSFET channel of the DEPFET pixel. The power spectral density of 1/f noise can
by approximated as:
S( f ) ∝
1
,
fα
Equation 17
where α is usually close to 1. 1/f is a low frequency noise which can be efficiently
suppressed by proper filtering which will be described in Chapters 4.1.3 and 4.1.4.
Page 38/116
The DEPFET Particle Detector Systems and Their Characterization
4
DEVELOPMENT OF THE DEPFET MINI-MATRIX TEST
SYSTEM
A versatile measuring system called Mini-matrix system has been designed to
characterize small prototypes of the DEPFET with 48 active pixels. The test system is
composed of commercial and custom-made blocks, such as a PC with an 8-channel PCI
data acquisition card [ 28 ], FPGA (Field-Programmable Gate Array) control card,
current readout pre-amplifiers, switching circuit, microprocessor slow control USB
card and power supplies. Red or infra red lasers and a 3-axis positioning stage with
spatial resolution of 1.25 µm are integrated into the system. The DEPFET sensor is
installed in the test chamber equipped with active thermal control system (Figure 19).
Figure 19 – Left: Configuration of the Measurement Setup. Right: Inside of the Test
Chamber
Figure 20 displays the conceptual schematic of the test system. The system concept was
proposed and designed by the author of this thesis, as well as the custom-made readout
Page 39/116
The DEPFET Particle Detector Systems and Their Characterization
and steering electronics and slow control. Some tasks were carried out by others in the
frame of the DEPFET Collaboration. In particular, the FPGA card called X-Board 2, as
well as the FPGA firmware were designed by the Max-Planck Institute in Munich [ 29 ],
[ 30 ]. Software for system control and data acquisition (DAQ) was developed mainly by
C. Oswald [ 20 ] and P. Kodyš [ 31 ]. Three versions of the system called Mini-matrix
V1, V2 and V3 were designed, to improve the system properties or extend the readable
sensor area. In [ 32 ] feasibility study and system simulations of system V1 are
described in detail. Measurements described in this thesis were measured with system
version V2 which is described in Chapter 4.1. Differences of other versions and the
reasons for modifications will be discussed in Chapter 4.2. Some parts of electronics of
the Mini-matrix system were used and tested with the MEDIPIX and TIMEPIX [ 33 ]
on the stratospheric balloon flights [ 34 ].
Figure 20 - Block Diagram of the Test System
Page 40/116
The DEPFET Particle Detector Systems and Their Characterization
4.1
Test System Design
In this chapter the custom made electronics is going to be described. The version of the
system V2, which was mainly used for the DEPFET mini-matrices tests, is going to be
described first. The differences in previous version V1 and the new V3 version will be
disused in Chapter 4.2.
4.1.1
System Version V2
The system is interfacing directly DEPFET sensor and it has to read sensor output
current with high precision. Minimal chip-to-amplifier connections are critical in the
design, as well as minimal amount of active components in the signal way. Therefore,
the specially designed current readout circuit consists of eight parallel low noise transimpedance amplifiers followed by non-inverting amplifiers (see Figure 21). The parallel
topology of individual drain channel readout makes no need of additional multiplexing
and current memory cells, which allows keeping design simple. The input node of the
amplifier is tied directly to the drain channel of the matrix. A 39kΩ resistor at the input
works as a pedestal current pre-subtraction and can be adjusted individually for each
channel by subtraction voltage trimming. Virtual drain voltage defines the input node
potential and offset voltage is used for additional offset setting. The preamplifiers are
optimized for low noise as well as for the best response for the step input signal.
Figure 21 - Schematic of the Preamplifier
Page 41/116
The DEPFET Particle Detector Systems and Their Characterization
DEPFET operation requires clocked voltages at the MOSFET gate and clear electrodes
for a pixel readout addressing and clearing. The correct sequence at digital level is
generated by the FPGA card and can be configured via control software. The digital
sequences have to be converted to analogue steering pulses. These pulses are provided
by a circuit with 12 individual pulse generators (switchers). A schematic of one channel
is in Figure 22. The analogue part is isolated from the FPGA card via digital isolators
ADuM110. Pulse buffers are realized by FET operational amplifiers (AD8066) operated
rail-to-rail. Analogue part of the pulse generator is fully isolated from the rest of the
system and is floating, so output “HIGH” and “LOW” voltages can be defined in wide
positive as well as negative range. Figure 23 shows a photo of the Mini-matrix system
and Figure 24 left, shows a photo of the preamplifier card.
Figure 22 - Schematic of the Pulse Generator
The system contains also linear regulators, power supply filters, digital isolators and
trigger circuits. USB slow control card was developed for temperature monitoring and
control of additional relay switches (Figure 24, right) and was implemented into the
mini-matrix system. It is equipped with four miniature Pt1000 temperature sensors,
which can be glued on the DEPFET carrier for temperature monitoring.
Page 42/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 23 – Photo of the Mini-matrix System
Figure 24 – Left: Photo of the Preamplifier, Right: Photo of the Slow Control Card with
Miniature Temperature Sensors
Page 43/116
The DEPFET Particle Detector Systems and Their Characterization
4.1.2
Sensor Signal Acquisition
The readout signals from the DEPFET sensor are the current base. The source of the
MOSFET pixel is kept at the ground potential and the MOSFET’s drain is at - 5 V. The
output signal current has two components: a fixed pedestal current, Iped, which does
not change with charge in the internal gate, and a signal current, Isig, which is
proportional to the integrated charge according to Equation 9. The pedestal current is
partially subtracted at the analogue level at the inputs of the low-noise readout
amplifiers and also digitally in the DAQ software. The drain signal currents are read out
and amplified by 8 amplifiers in parallel and digitized by a GaGe Octopus data
acquisition card with 8 14-bit inputs.
Figure 25 - Scope Plots of the Rolling Shutter Readout Scheme
Page 44/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 26 – Drain Current Scope Plot
In Figure 25 is depicted the “rolling shutter” readout scheme of the DEPFET in gate
and clear channels. Figure 26 shows a typical digitized stream of the DEPFET matrix
drain channel. One pixel is determined by corresponding drain and gate channel. It
means, one frame consists of 6 gate and 8 drain channels (48 pixels). The signal is
inverted by preamplifiers, so more negative values means higher signal. Positive spikes
in the drain readout stream are caused by clearing pulse. MOSFET channel has
capacitive coupling to the clear contact and clear pulse is than visible in the drain signal
during the clear process. Switching the gates can also cause spikes in the drain current.
These spikes can be efficiently suppressed be adjusting overlaps of gate pulses. The
signal of each pixel is evaluated according the following scheme:
1.
One row of the pixels is turned on by applying negative voltage at the FET gate.
2.
Charge is integrated in the internal gate.
3.
Drain current of each transistor in one row is measured in parallel.
4.
The whole row is cleared.
5.
Drain current is read again.
6.
The row is turned off, following row is turned on and the process is repeated.
Page 45/116
The DEPFET Particle Detector Systems and Their Characterization
The signal is evaluated as a difference of currents after clear (AC) and before clear (BC).
This readout scheme is called correlated double sampling (CDS) [ 35 ]. CDS method is
combined with averaging of multiple samples in the Mini-matrix system. This filtering
method known as a multi-CDS [ 36 ] represents an improvement in compare with
traditional RC-CR shapers, widely used in particle physics experiments. Impact of
averaging and CDS on noise performance will be discussed in the following chapters.
10
9.5
9
Noise [ADU]
8.5
8
7.5
7
6.5
6
5.5
5
0
20
40
60
80
100
Number of Samples [ - ]
Figure 27 – RMS Noise versus Number of Samples in One Acquisition Point
4.1.3
Averaging
The acquisition points AC and BC are not just single measurements, but several
consecutive samples which are averaged. The standard deviation 0f the averaged
( )
samples σ i AC is given by:
( )
σ i AC =
σ (i AC )
,
Equation 18
n
where σ (i AC ) is a standard deviation of one sample measurement and n is number of
samples. The same is valid for samples before clear (BC). Figure 27 displays results of
Page 46/116
The DEPFET Particle Detector Systems and Their Characterization
RMS noise (1σ noise) measurement versus number of samples in one acquisition point.
It is in agreement with Equation 18. Since σ(iAC) is constant for each sample point the
RMS noise of averaged n samples decreases with 1/√n. Averaging can significantly
improve system noise, but it is limited and averaging over too many samples will not
improve noise performance. Typically number of samples is selected n = 30 in the
Mini-matrix system.
4.1.4
Correlated Double Sampling
Correlated double sampling (CDS) was introduced by White [ 35 ] in 1974 for noise
elimination in charge coupled devices (CCDs). Today, this method is widely used in
imaging detectors.
Signal subtraction of two acquired values i BC (before clear) and i AC (after clear) is used
in correlated double sampling. A time difference between two samples is labelled as τ.
Since i AC is a pedestal current and i BC is a signal current plus pedestal current, when
these values are subtracted the slow variations of the pedestal current are suppressed.
In a frequency domain, a noise transfer function of CDS HCDS is given by [ 37 ]:
H CDS (ω ) = 1 − e − jωτ ,
Equation 19
where j is a complex unit and ω is an angular frequency. Square modulus of HCDS can be
expressed as:
H CDS (ω ) = 2 ⋅ (1 − cos(ωτ )) .
2
Equation 20
Square modulus of the CDS noise transfer function is illustrated in Figure 28. Multiples
of 2π/τ are completely cancelled, but intermediate frequencies are even multiplied by
factor of two. The CDS offers good filtering of low frequencies, but it is necessary to
Page 47/116
The DEPFET Particle Detector Systems and Their Characterization
introduce a low-pass filter to the system to achieve a finite RMS noise as illustrated in
Figure 28. The low-pass filter is realized as a one pole filter at analogue level in the
Mini-matrix system. Square modulus of low-pass filter HLPF is given by
H LPF (ω ) =
1
2
 ω
1 + 
 ωC



2
,
Equation 21
where ωC is a cut-off frequency.
Figure 28 – Square Modulus of CDS, One Pole Low-pass Filter and a System of CDS and
Low-pass Filter Noise Transfer Functions
Different noise sources are present in the DEPFET sensor, as thermal noise, 1/f noise
and shoot noise. Impact of CDS on noise suppression of different noise sources was
observed. Total power spectral density s(f) of this three noise sources can be expressed
as:
s ( f ) = sth +
2
s12/ f
f
α
2
+ sshot
,
Equation 22
where sth, s1/f and sshot are power spectral densities of thermal, 1/f and shot noise, f is
frequency. 0 < α < 2 and α is usually close to 1. In case of the thermal noise the spectral
Page 48/116
The DEPFET Particle Detector Systems and Their Characterization
density is frequency independent (white noise). The thermal noise suppression is
mainly caused by bandwidth limitation of the low-pass filter. In case of 1/f noise the
CDS method become very effective, because of high noise spectral density at low
frequencies. In contrast with RC-CR shaping filters, which have similar transfer
function for noise as well as signal, the CDS filtering method offers different transfer
functions for uncorrelated (signal) and correlated (noise) inputs [ 36 ]. SNR might
significantly increase, when CDS is applied. A leakage current in the detector generates
a signal offset during the integration time. Random fluctuation of the leakage current
causes noise contribution known as a shot noise. During the CDS process the device is
reset (cleared) to sample a reference baseline signal. During the reset also shot noise
contribution is cleared and there is no correlation between these two samples.
Therefore, shot noise cannot be suppressed by the CDS. Figure 29 shows results of
RMS noise measurement on the Mini-matrix setup and illustrates noise suppression by
CDS filtering method.
12
Noise [ADU]
10
8
6
4
2
0
0
5
10
15
20
Correlated Double Samples Distance [µs]
Figure 29 – RMS Noise versus Correlated Double Samples Time Difference
When single sampling is used, the clear pulse is applied only after the readout of
several frames and the drain current increment provides the measurement of the
signal. The correlated double sampling is mainly used in the Mini-matrix system
because of the high 1/f noise immunity. On the other hand, the complete clear is
Page 49/116
The DEPFET Particle Detector Systems and Their Characterization
critical, when correlated double sampling scheme is used. Incomplete clear will
introduce additional noise known as a reset noise.
The signal stream is recorded and acquired by the data acquisition system (DAQ). The
DAQ have to read the raw data and subtract signals of each pixel.
4.1.5
Software Package
A software package which controls the Mini-matrix system called MiMaTools was
developed by C. Oswald and P. Kodyš. Detailed description can be found in [ 20 ]. Since
it was developed in close touch with the author of this thesis and it is important part of
the system, brief description will be listed.
Figure 30 – Overview of the MiMaTools Software Package - C. Oswald [
20 ]
The software package is divided in three main parts. Hardware control part
communicates the MinaTools software controls the power supplies and moves the
Page 50/116
The DEPFET Particle Detector Systems and Their Characterization
optical laser focuser. Data acquisition software (DAQ) runs data taking of three basic
measurements – noise, radioactive source, or laser measurements. Analysis software
than interprets the data sets in a graphical form.
4.1.6
Test Environment
The Mini-matrix test system is implemented into a test environment, which can control
temperature and illuminate the DEPFET sensor with a laser beam at defined place.
Temperature control system is based on a liquid cryo-circulator connected to a liquidto-air heat exchanger which is placed inside of the test chamber. A feedback
temperature sensor of a PID regulator is placed inside of the test chamber as well. The
inside of the test chamber is insulated by a styrene-foam and metalized milar blanket to
minimize the thermal resistance between the test chamber and ambient environment.
Figure 31 illustrates thermal stability inside and outside the test chamber over 24
hours.
Figure 31 – Temperature Measurement Outside and Inside of the Test Chamber
Page 51/116
The DEPFET Particle Detector Systems and Their Characterization
Two laser modules with deep modulation 660 and 1060 nm developed in CERN are
used for illumination of the DEPFET sensor. These lasers are connected via an optical
fibre to a focuser which is attached to a 3-axis position stage with spatial resolution of
1.25 µm. Figure 32 shows a photo of the positioning stage with the focuser and a
mechanical support frame for the Mini-matrix setup. Measurements with radioactive
sources had been done by placing the radioactive source above the DEPFET sensor.
Figure 32 – Photo of Laser Optics and Positioning Stage
4.1.7
Test System Characterization
The system has been characterized and tested with the DEPFET mini-matrices of the
PXD5 production. The characterization was mainly done by C. Oswald within his
diploma thesis [ 20 ]. First part of the characterization investigated shape of the
switching signals and their stability. Especially, stability of the steering pulses and the
same shape of the pulses from different channels are critical for the operation of the
system. A difference in amplitudes of the gate pulses causes different operational point
setting of each matrix row, as well as different pedestal current. Ringing and
Page 52/116
The DEPFET Particle Detector Systems and Their Characterization
oscillations in gate pulses are mirrored in the drain current and makes sampling
difficult. Another important issue is a stability of rising and falling edges of the gate
pulses. Overlaps of consecutive gate pulses are used to achieve almost continuous drain
current flow. Differences in falling and rising edges of different gate channels makes
necessary individual adjustment of the gate pulses overlap. It leads to complicated
sequences which may be used on specific sensor only.
Instability of the steering pulses of the system version V1 was the main reason for
development of the second system V2. Figure 33 shows measurement of gate pulse
made by C. Oswald. Left figure shows overlay of all 6 gate channels and it illustrates
similarity of all gate pulses. Right figure shows zoom of one gate pulse. Two spikes in
the middle of the gate pulse are caused by a crosstalk with the clear pulse. Figure 34
shows the overlay of the falling edges of the gate pulses. The time shift of different
pulses is smaller than 2 ns (time resolution of the digital sequencer is 7.5 ns). This
allows high precision configuration of the overlap of the consecutive gate pulses and
fulfil the system requirements.
Figure 33 – Left: Gate Pulses (6 channels overlay). Right: Gate Pulse Zoom (1 channel).
C. Oswald [
20 ]
Page 53/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 34 – Falling Edges of the Gate Pulses. C. Oswald [
20 ]
There are eight pre-amplifiers in the Mini-matrix system and due to the tolerances of
components used in these amplifiers there are some small differences in gains. The
average transconductance is 8200 nA/V, but there are differences among the amplifiers
in range of 80 nA/V. Each amplifier has been calibrated and its value is stored in the
configuration file of the DAQ. Figure 35 shows transfer characteristic of one drain
channel and residual of a linear fit (red). Small deviation of the input current from
linear fit (Figure 36) was observed, when output voltage crosses 0 V level. It is probably
caused by construction of used operational amplifier. Operation range was than shifted
to negative values of the output voltage to prevent crossing of 0 V level.
Figure 35 – Left: Transfer Characteristic of the Amplification Chain. Right: Residuals of a
Linear Fit. C. Oswald [
20 ]
Page 54/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 36 – Deviation of the Input Current from a Linear Fit versus Output Voltage.
C. Oswald [
20 ]
Figure 37 – Left: Stability of the Drain Current (Red: After Clear, Blue: Before Clear),
Right: Stability of the Signal - C. Oswald [
20 ]
Figure 37 presents the stability of the signal in long term duration. Left, the stability of
the drain current is depicted. (The thermal control wasn’t installed yet.) One can see a
drift of drain current in time, caused by heating up of the system. Right, the stability of
the signal is shown in 24 hours. One can see, that signal is stable, even if the drain
current is changing. The noise performance of the system was tested mainly with a real
DEPFET detector will be discussed in section 6.
Page 55/116
The DEPFET Particle Detector Systems and Their Characterization
4.2
Discussion of the Test System Design
Development of the Mini-matrix test system started in 2007 and in 2009 first version
of the system Mini-matrix V1 was finished. The basic concept was the same as in the
system V2, but the preamplifiers had 3 stages. The last stage was a differential buffer,
which could be configured to drive differential or single ended inputs. The differential
option was never used and the buffer was removed in the next version of the system.
The system V1 has also SPI (Serial Peripheral Interface) multi channel DAC (Digital-toAnalog Converter) for subtraction voltage adjustment. This feature was replaced in the
future versions by manual screw trimmers. The most important change in the following
designs was redesign of the pulse generators (switchers). In system V1 an analog
CMOS switches were used (see Figure 38). These switches had complicated powering
and grounding which leaded to ringing and cross talks in the switcher channels. Figure
39 illustrates improvement of the new switcher design.
Figure 38 - Schematic of the Old Pulse Generator
Page 56/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 39 – Scope Plots of the Pulse Generators, Left: V1, Right: V2
Three copies of The Mini-matrix test system V2 were manufactured and two of them
were supplied to Max-Planck Institute in Munich and Ludvig-Maximilians University
in Munich. The systems were primarily operated with 8-channel ADC cards. For costsaving reasons the system was also equipped only with 4-channel ADC card in Prague
laboratory. For this reason a multiplexer board was installed to the system, which
allows switching between two sets of four channels by a command via slow control USB
card. Figure 40 shows, how the multiplexer card is installed in the system.
Figure 40 – Installation of the Multiplexer Board
When PXD6 DEPFET generation detectors were introduced, the 4-fold readout layout
(see Figure 17 and Figure 18) leaded to low amount of active columns of the detector. In
the 4-fold readout scheme, only 2 columns are active, when 8 drain channels are
Page 57/116
The DEPFET Particle Detector Systems and Their Characterization
connected. It is insufficient for some required measurements so, development of new
version - Mini-matrix V3 – was started (see Figure 41).
Figure 41 – Assembling of Mini-matrix V3 Main and Amplifiers Boards
The Mini-matrix V3 has a new main board, which is compatible with the V2 version.
The size of the new PCB is identical and all the connectors are at the same positions.
The difference is in amount of the DEPFET drain channels, which can be connected.
The new system has 16 parallel channels. The amplifier cards are more than thee times
smaller, than the previous one, so they will fit on the new main board. The system uses
analog multiplexers for switching between two sets drain channels. One X-Board
trigger channel is used for multiplexer control. Two sets of samples before clear (BC)
and after clear (AC) will be shifted in time. The principal schematic of operation is
shown in Figure 42. One half of the detector will be sampled before clear and than the
multiplexer will be switched and other half of the detector will be sampled before clear.
Than the clear pulse will be applied and two sets of samples after clear will be acquired.
The DEPFET sensor itself will be used as a memory unit in this design and none
external memory cell will not be required. Such design, will allow using the same XPage 58/116
The DEPFET Particle Detector Systems and Their Characterization
Board firmware and ADC card as in the precious design. Slightly higher noise of the
new system is expected due to the more complicated design. There will be small
difference in integration time in each parts of detector, so small difference in pedestal is
expected. It is expected, that the new system will be more suitable for new detector
designs.
Figure 42 – Principles of Signal Multiplexing
Page 59/116
The DEPFET Particle Detector Systems and Their Characterization
5
MEASUREMENTS
There are several kinds of measurements that may be performed at the Mini-matrix
test system. Types of the measurements and measurement methods will be described in
this chapter. The system was mainly used to test detectors with laser beam or just to
evaluate noise of the detector at different operational voltages. Other possibility is
usage of the radioactive source.
5.1
Data Analysis Steps
The analogue data are digitized and stored on a hard disc as described in Chapter 4.1.1
System Version V2. The following steps are applied on the data sets in the off-line
analysis chain.
Removing of pedestals
The pedestals are remaining signals of each pixel, which
were acquired by detector without any signal (in the dark). The pedestals are caused by
leakage current or parasitic effects of steering electronics. Several thousands of these
pedestal frames are acquired and then the mean value of pedestal for each pixel is
subtracted from the signal in the further analysis.
Common mode correction
The correlated variation of the pixel signal can be
observer during the measurement. This variation is usually caused by electronic
systems, which controls the detector and readout amplifiers. Especially, instability of
gate pulses or changes of system voltages can cause common change of the detector
response. The main part of common mode noise is filtered out by the correlated double
sampling as was described in Chapter 4.1.4 Correlated Double Sampling, but small
fraction of this kind of noise is still present in the data and the noise performance can
be improved, when the variation of the common pixel signal is corrected. Deeper study
of common mode correction at the Mini-matrix system is shown in [ 20 ].
Page 60/116
The DEPFET Particle Detector Systems and Their Characterization
Clustering Events (particle or laser hits) acquired by the pixel detectors are usually
spread over several pixels. This is very important for high spatial resolution of the pixel
detector. When charge is distributed over multiple pixels, position of the hit can be
calculated by using the centre of gravity algorithm [ 38 ] with higher resolution than
pixel pitch. This parameter was carefully tested [ 38 ] to optimize detector thickness.
When charge of one event needs to be calculated, it is necessary to sum charge over all
pixels affected by the hit. The pixels with signal higher than 6σ of noise level are
marked as seeds in this process. The pixels with signal higher than 2σ of noise level
are added to a cluster. The thresholds levels for seeds and clusters can by tuned and a
cluster size is also usually limited. The maximal cluster size is set to 3 x 3 pixels in the
Mini-matrix system due to the limited amount of active pixels.
5.2
Noise Measurements
Several kinds of noise can be observed: intrinsic noise of the DEPFET and electronic
noise. Intrinsic noise of the DEPFET is quite low and it is mainly dependent on
bandwidth of the system. In [ 39 ], the measurement of the DEPFET (PXD4
production) intrinsic noise is described. The intrinsic ENC noise 2.1 electrons was
measured during spectroscopic measurement with
55Fe
source using RC-CR shaping
amplifier [ 39 ]. Since the DEPFET contribution to the ENC noise is small in compare
to the ENC noise of the electronic which is approximately 20 electrons, any tendencies
in the ENC noise measurement are mainly system related. The system noise is
evaluated as a standard deviation of the pedestal signal (RMS). Measurements were
proceeded with the running detector in the dark without any radioactive source.
5.3
Radioactive Source Measurements
Radioactive source measurements can be used for detector energy calibration.
Calibration of the detector can be done with particles or photons with known energy
Page 61/116
The DEPFET Particle Detector Systems and Their Characterization
deposition. It can be achieved by placing the detector into the accelerator’s beam, but it
requires a lot of resources. More about tests of DEPFET on the CERN SPS accelerator
can be found in [ 38 ]. Another way how to calibrate the detector is by using a
radioactive source. Detector tests with x-ray and gamma radioactive sources are
convenient, because the energy spectrum of the source is known and it gives us
information about the detector performance in a way more similar to the real operation
of the sensor than the laser test. Measurements with the radioactive source gives exact
knowledge of the deposited energy and therefore also created charge. The source tests
are widely used for detector calibrations. However, the sensitive area of the minimatrix detector is very small and the system readout is relatively slow (usually 20 to
200 µs / frame), acquisition of a single spectrum takes at least several hours.
The easiest method to acquire radioactive source spectra is to sum charge over all pixels
which is higher than certain threshold level and draw histogram of total deposited
charge. Such method includes also incomplete clusters at the edges of the sensor, but it
has high efficiency and it is the fastest method. Different techniques can be used to
improve the selection of hits. The commonly used method is clustering with additional
constrains (cuts) and hit recognition algorithm, which selects cluster created by
particles. The cuts will eliminate unwonted effects. These might be hits with seed pixels
at the edge of the sensor (incomplete cluster), or single pixel clusters, which are usually
random noise. The cuts have to be carefully selected otherwise they can distort the
measurement. The integrity of the clustering method can be verified by a correlation
plot. An example of the correlation plot measured at Mini-matrix system is shown in
Figure 43. The x-axis shows a number of a pixel, which is a seed pixel in a cluster. The
y-axis shows pixel number of all pixels, with a signal. When the sensor is bombarded by
particle from a radioactive source, statistically, neighbouring pixels to the seed pixel
should be hit more often, than other pixels. Obviously, seed pixels will always appear in
“All pixel hits” on y-axis. When the clustering works well, the diagonal band will appear
in the correlation plot. Due to the limited amount of active pixels the Mini-matrix
system is not an optimal tool for spectroscopic measurements. The active area is very
small and due to the “rolling shutter” readout, the clusters can be split to two
consecutive frames and not recognized as a single cluster. Especially, layout of PXD6
Page 62/116
The DEPFET Particle Detector Systems and Their Characterization
DEPFET generation, which has 2 x 24 active physical pixels, does not allow clustering
at all.
Figure 43 – Example of Correlations of Seed Pixel Hits vs. All Pixels Hits
5.4
Laser Measurements
Laser tests of the semiconductor detectors are quick and efficient way how to evaluate
the properties of the semiconductor detectors and test systems. Especially, spatial
response of the detector can be investigated with the laser beam. In contrast with the
radioactive source, where hits are random in time and space, laser pulses are
synchronized with the readout and the position of the laser focuser controlled by the
system. The sensors are always illuminated from the back side where is an opening in
the metaization. Front surface is covered by metal lines, which would not allow
penetration of light.
Two kinds of lasers were used with the Mini-matrix system, red, 660 nm, laser and
infrared (IR), 1060 nm. According to Figure 11 the absorption length of 660 nm photon
Page 63/116
The DEPFET Particle Detector Systems and Their Characterization
is few micrometers in silicone. It means, that the red laser photon penetrates through
pasivation surface layer and reaches depleted region, where it is absorbed by photo
effect with 100 % probability. Red light is than advantageous due to the homogeneous
response in space of the sensor. But full depletion of the sensor is crucial for detection
of the red light. Absorption length of 1060 nm photon is around 800 µm. It is more
than thickness of the sensor and some fraction of the laser pulse energy is not absorbed
by silicone. The IR laser beam generates electron-hole pairs all its way through the
silicone and it is more similar to the particles. On the other hand, reflections on the
front surface may occur and spatial response of the detector scan with the IR laser
beam is affected by reflections.
The focused laser beam has a Gaussian distribution with σ = 3 µm. Since the charge
generated in the silicone diffuses in lateral way and the pixel is not a single point, but
an area, the single pixel sees laser beam as shown in Figure 44.
Laser pulse is
synchronized with the readout and laser is fired on the beginning of each frame.
Position of the trigger with respect to the rolling shutter readout scheme is depicted in
Figure 45.
Figure 44 – Laser Spot Size
Several kinds of laser measurement can be performed on the Mini-matrix system. The
fastest method is a single pixel measurement. In such measurement, the laser
beam is pointed just to a single point and signal of the hit pixel is being readout. It can
be used especially for voltage scans and optimization of the operational point of the
Page 64/116
The DEPFET Particle Detector Systems and Their Characterization
sensor. Full matrix surface scan is usually used for investigation of spatial response
of the sensor. It can be also used with combination of voltage scan, but the
measurement will take much more time. The laser beam scans the surface of the
sensor with predefined step (min. 1.25 µm) and the sensor is read out several times in
each point.
Figure 45 – Laser Trigger Synchronization with the Readout
A special kind of the laser measurement which was carried out using the Mini-matrix
system is gated mode of the DEPFET. The gated operation makes possible to blind
(make insensitive) sensor for incoming radiation. A special steering sequence shown in
Figure 46 was applied to the DEPFET matrix. It has an additional shielding clear pulse,
which is added to the standard sequence to emulate the insensitive period. The laser
trigger was applied once during the frame readout, in the beginning of the frame, or
during the insensitive period. The first laser trigger is used to generate charge before
the blind mode is applied and to evaluate, if there are some charge losses due to the
additional shielding clear pulse. The second laser trigger (applied during the insensitive
mode) emulates the junk noisy charge, injected by noisy bunches of the accelerator and
it allows to evaluate the junk charge selection.
Page 65/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 46 – Blind Mode Sequence
Page 66/116
The DEPFET Particle Detector Systems and Their Characterization
6
RESULTS
6.1
Electrical Characteristics
As a first step of a characterisation of the DEPFET device an optimal voltage operation
point has to be found. Source of the MOSFET in the DEPFET pixel is tied to analogue
ground AGND and all other DEPFET related voltages are referred to source (AGND).
The voltages can be divided in several categories according to the purpose of the voltage
and a structure, where it is applied.
6.1.1
MOSFET Related Characteristics
First group of voltages are related to the MOSFETs. These are drain voltage and gate
voltage. Gate voltage is clocked and it has two levels: Gate_LOW and Gate_HIGH,
which corresponds to on- and off-state of the MOSTET. In the on-state, the MOSFET is
operated in saturation region (Gate_LOW < Vth , where Vth is a threshold voltage). The
MOSFET channel is pitched-off and lack of channel region near the drain causes weak
dependency of drain current on drain voltage. The drain current is mainly controlled by
Gate_LOW voltage and can be modelled approximately as [ 40 ]:
Id =
1
W
2
µ p Cox′ (Vgs − Vth ) (1 + λVds ) ,
2
L
Equation 23
where is µp is a charge carrier mobility, Cox’ is a sheet capacitance and W and L is a
width and length of the gate, λ is a channel-length modulation parameter and Vds is the
drain-to-source voltage. The λ parameter is responsible for the differential output
resistance of the DEPFET pixel. Output resistance rout can be described as:
Page 67/116
The DEPFET Particle Detector Systems and Their Characterization
rout =
∂Vds
1
=
.
∂I d
λVds
Equation 24
Figure 47 shows a measurement of drain current Id vs. drain-to-source voltage Vds. The
Mini-matrix system does not allow measurement of absolute drain current, but only
relative changes so the drain current scale is relative only. The differential output
resistance of the DEPFET pixel cell is rout = 1.42 MΩ. The drain voltage is set to -5 V.
Drain current [µA]
2.5
2
1.5
1
0.5
0
-6.5
-6
-5.5
-5
-4.5
-4
-3.5
-3
Drain-to-source voltage [V]
Figure 47 – Drain Current vs. Drain-to-source Voltage
Charge in the internal gate has similar effect on the drain current as a gate voltage
change. When charge in the internal gate is present, the Equation 23 can be modified as
[ 41 ]:
1
′W
I d = µ p Cox
2
L
 q
 f
+ Vgs − Vth
 Cox
2

 (1 + λVds ) ,

Equation 25
where q is the charge in the internal gate, Cox is a capacitance of the gate and f is a
reduction factor due to the parasitic coupling of the internal gate. For further
calculations the λ can be neglected.
According to Lutz [ 41 ], when DEPFET is operated in saturation region the
transconductance gm of the DEPFET pixel can by calculated as:
Page 68/116
The DEPFET Particle Detector Systems and Their Characterization
gm =
∂I d
′W
(Vgs − Vth )2 .
= − µ p Cox
∂Vgs
L
Equation 26
Combining Equation 23 (λ is neglected) and Equation 26 results in:
g m = 2µ p Cox
′W
Id .
L
Equation 27
Measurement of drain current vs. Gate_LOW voltage is shown in Figure 48. According
to Equation 26 transconductance is calculated: gm = -26.4 µA/V.
When the Gate_LOW voltage is close to 0 V, there is still a drain-to-source leakage
current flowing. Since the Mini-matrix setup can not measure absolute drain current
values also absolute value of the drain-to-source leakage current cannot be measured.
On the other hand the shape of the IV-characteristics and the relative leakage current
behaviour is visible in a detail of the sub-threshold leakage drain current shown in
Figure 49. The leakage current is caused by thermal ionization, amplification of gateto-channel band-to-band tunnelling current or gate-to-drain tunnelling current [ 42 ].
The theory of tunnelling [ 43 ] predicts behaviour of the leakage current Id according to
the following formula which fits to Figure 49:
I d = − AE S e − B ES ,
where A and B are experimental coefficients and ES
Equation 28
is an intensity of a vertical
electrical field at silicone surface.
The internal amplification gq is proportional to the transonductance gm as:
gq =
gm
.
Cox
Equation 29
Page 69/116
The DEPFET Particle Detector Systems and Their Characterization
One can see that the value of the gq is related to the parameters of the MOSFET as L,
W, µp and Cox’. These parameters are fixed during the design of the sensor and only way
how to adjust the gq during the operation of the sensor is variation of the drain current
Id. Relation between gq and Id is approximately:
gq ≈ Id .
Equation 30
Measurements of seed response vs. Gate_LOW voltage of three detector types are
shown in Figure 50. The signal charge, generated by laser is constant during the
measurement, so the changes of the signal current are caused by variation of the gq.
Since gq is proportional to the Id as a square root Figure 50 is in agreement with
Equation 30. The Gate_LOW voltage is set to -4 V to achieve highest gq and reasonably
Drain current [µA]
low drain current.
90
80
70
60
50
40
30
20
10
0
-10
-6
-5
-4
-3
-2
-1
0
Gate_LOW [V]
Figure 48 – Pedestal Drain Current vs. Gate_LOW Voltage – PXD6-50µm
Page 70/116
The DEPFET Particle Detector Systems and Their Characterization
Drain current [µA]
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
-0.5
-2
-1.5
-1
-0.5
0
Gate_LOW [V]
Figure 49 – Sub-threshold Pedestal Drain Current vs. Gate_LOW Voltage – PXD6-50µm
Signal current [nA]
4000
3500
3000
2500
PXD5
2000
1500
PXD6-450µm
1000
PXD6-50µm
500
-8
-6
-4
-2
0
Gate_LOW [V]
Figure 50 – Seed Response vs. Gate_LOW Voltage
6.1.2
Clear Related Characteristics
Another set of voltages are related to the clearing process and clearing structures.
These are Clear_HIGH, Clear_LOW and Clear gate voltages. Clearing performance is
Page 71/116
The DEPFET Particle Detector Systems and Their Characterization
important, since noise and signal performance is depending on it. Clear_HIGH voltage
is applied to the clear electrode during the clearing process and its amplitude is crucial
for complete clear and clearing efficiency of the device. Clearing efficiency of PXD5 and
PXD6 detector families can be seen in Figure 51. Inefficient clearing is observed as
rising of the pedestal drain current. Due to the reset noise of the correlated double
sampling it is necessary to keep clearing efficiency high. The Clear_HIGH voltage is
kept around 19 V for all production series.
8
7
PXD6-450µm
Drain current [µA]
6
PXD5
5
PXD6-50µm
4
3
2
1
0
-1
5
10
15
20
25
Clear_HIGH [V]
Figure 51 – Clear Efficiency vs. Clear_HIGH Voltage
The Clear_LOW voltage defines base line of the clear pulse. It cannot go too negative,
otherwise electrons are back injected from clear contact and integrated in the internal
gate. Too high value will cause loosing of electrons from the internal gate during
integration and sensor would be continuously cleared. Figure 52 shows effect of
Clear_LOW voltage on pixel signal. Clear_LOW is kept around 3.8 V. Common clear
gate contact is introduced to PXD5 and PXD6 production series as described in Chapter
3.2. Static clear gate voltage helps to reduce Clear_HIGH voltage and has to be
carefully tuned with Clear_LOW voltage. Signal current modulation vs. Common clear
gate voltage is depicted in Figure 53. Some parasitic effect can occur at different
Page 72/116
The DEPFET Particle Detector Systems and Their Characterization
combinations of Clear_LOW and Common clear gate voltage. These effects are
illustrated in Figure 54, where both voltages were swept in 3D plot.
12000
PXD6-50µm
Signal current [nA]
10000
PXD5
8000
6000
4000
2000
0
0
2
4
6
8
10
Clear_LOW [V]
Figure 52 – Seed Response vs. Clear_LOW Voltage
3500
Signal current [nA]
3000
2500
2000
PXD6-50µm
1500
PXD6-450µm
1000
PXD5
500
0
-4
-3
-2
-1
0
Common clear gate [V]
Figure 53 – Seed Response vs. Common Clear Gate Voltage
Page 73/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 54 – Seed Response vs. Gate_LOW and Common Clear Gate Voltage of PXD5
Detector
Following parasitic effects were observed:
1. Back Emission: Electrons are emitted from the clear implantation and
integrated in the internal gate, when Clear_LOW voltage is too low. Back
electron emission leads to detector saturation.
2. Charge loss: Too high Clear_LOW voltage causes continuous clearing and lost
of charge from the internal gate. Too positive common clear gate voltage
causes accumulation of electrons under the clear gate contact which will not
reach the internal gate and will be lost during the clearing cycle.
3. Hole inversion: When the clear gate voltage is too negative a p-channel can
develop in the n-bulk under the clear gate contact. This channel can short
drain and source or source and clear p-well and will disable sensor operation.
Page 74/116
The DEPFET Particle Detector Systems and Their Characterization
6.1.3
Other Characteristics
The remaining set of voltages is bias back depletion voltage, bulk voltage and drift
voltage. Bulk voltage will be discussed in the following chapter separately. Drift voltage
is introduced only to the PXD6 detector families and will be discussed later too.
Depletion voltage is connected to the back contact and it is depleting back side of the
detector. Figure 55 shows a measurement of a seed charge versus back depletion
voltage. A sharp threshold is visible in measured plots. Detector is not fully depleted
below threshold voltage and it is not sensitive for light. Since voltage measurements
were done with the red laser, charge is generated only at the back surface of the
detector and full depletion is necessary for signal detection. 450 µm thick detectors are
fully depleted around -120 V. 50 µm thick detector is fully depleted around -21 V.
3500
Signal current [nA]
3000
2500
2000
1500
PXD6-450µm
1000
PXD6-50µm
500
0
-220
PXD5
-200
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
Depletion voltage [V]
Figure 55 – Seed Response vs. Back Depletion Voltage
6.1.4
Edge Effect
Pixels, which are located at the edges of the detector, might suffer by lost of charge,
which escapes to bulk and surrounding edge structures.
The Mini-matrix system
usually does not use whole area of the detector, so at some detector samples the edges
Page 75/116
The DEPFET Particle Detector Systems and Their Characterization
of the detector are not accessible (all PXD6). PXD5 mini-matrices tested at the system
allow access to two edges of the detector. Figure 56 show a 2D laser scan of the PXD5
detector. Direction in x axe is row-wise. Direction in z axe is column-wise. First and last
rows (right and left in Figure 56) are the edges of the detector and one can see that first
and last two rows are affected by the edges of the detector. Variation of the bulk voltage
improves charge loss at the edges.
Figure 56 – Seed Response vs. Laser Position
Bulk voltage contact is located at the outer ring of the sensor. Bulk voltage has to be
positive with respect to source otherwise bulk-source PN junction become polarized
forward and current is injected from bulk to MOSFET channel. Figure 57 presents a set
of 2D laser scans with cluster charge maps for various bulk voltages. When bulk voltage
is rather positive, electrons generated at the edge of the detector by laser are attracted
by bulk ring and edge pixels become insensitive. Beginning of bulk current injection is
visible, when bulk voltage is 0.5 V. Pixels, where bulk current injection occurs are
completely saturated and read as zero. Optimal bulk voltage was found at 2 V.
Additional edge rings were placed at some DEPFET prototypes, which gives more space
for optimization, but these rings were not connected in the Mini-matrix system.
Page 76/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 57 – Cluster Charge vs. Laser Position
6.1.5
Drift Regions
The Belle II design PXD6 matrices which have drift regions added to the larger pixels
have additional drift voltage which has to be optimized. It improves drift of electrons
from the area of the enlarged pixel. The voltage applied to the drift regions has strong
effect on the charge collection of the detector. Too positive drift voltage causes
inefficient charge collection. Figure 58 shows a measurement of seed charge versus
drift voltage. Poor charge collection was observed for higher drift voltages than -1 V. It
is caused by low lateral electrical field in the pixel and very slow electron drift to the
internal gate. When the electrons do not reach the internal gate in the integration time,
the clear cycle is performed and the charge is lost. The single point measurement
shown in Figure 58 gives rough estimation about relation between the drift voltage and
charge collection, but since in this case, the charge collection is highly dependent on
position, where the charge is generated, 2D laser scans gives better view about charge
collection process.
Page 77/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 59 presents a set of 2D cluster charge measurements for different drift voltages.
When the drift voltage is close to 0 V, charge collection is poor in regions more remote
from the centres of the quadruple pixels with common gate electrodes (see Figure 18).
Black contours show the measurement of seed charge collection at optimal operational
voltages. It indicates the centres of the pixels. The charge collection is homogeneous for
drift voltage higher than 1 V. More negative drift voltage negatively affects gain of the
detector. Drift voltage was set to -2 V.
4000
Signal current [nA]
3500
3000
2500
2000
1500
1000
500
0
-10
-8
-6
-4
-2
0
Drift voltage [V]
Figure 58 – Seed Response vs. Drift Voltage – PXD6-450µm
Page 78/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 59 – Cluster Charge vs. Laser Position
6.1.6
Discussion of the Electrical Characteristics
Definition of electrical characteristics and their optimization is important for well
understanding of the behaviour of the detector. In the previous chapters are presented
some differences among different senor prototypes, but small characteristic deviations
were observed also among sensors of the same type. Limitations of the final ASICs also
were not taken into account and especially high clear voltage, which shows good clear
Page 79/116
The DEPFET Particle Detector Systems and Their Characterization
efficiency, might be pushed down by need of low power consumption of switcher
ASICs.
Another important issue is a change of MOSFET threshold voltage in the DEPFET pixel
due to the irradiation. Figure 60 measured by S. Rummel [ 39 ] shows results of
DEPFET irradiation and shift of threshold voltage which has to be compensated by
change of gate voltage. In the Belle II experiment, where high radiation background is
expected, the detector has to be able to compensate such shift and work properly.
Figure 60 – Threshold Voltage Shift Caused by Irradiation [ 39 ]
6.2
System Noise
The noise of the DEPFET detector and DEPFET sensor itself are closely tied together.
The overall noise value is given by intrinsic noise of the DEPFET sensor, noise of all
amplifiers in the signal way, but also by noise of the power supplies and voltage
regulators and noise of the pulse generators. It would be difficult to separate these
noise sources and measure noise contribution of each of them. So the system noise was
measured which defines noise of the DEPFET Mini-matrix system including all
electronic subsystems. The system noise also depends on the timing of the steering
Page 80/116
The DEPFET Particle Detector Systems and Their Characterization
sequences and averaging and correlated double sampling adjustment as was discussed
in Chapter 4.1.3 and 4.1.4. Figure 61 shows a noise histogram of one pixel pedestals of
the PXD6 DEPFET matrix. RMS noise 3.9 ADU corresponds to 7 nA of equivalent
input referred noise current or 20 electrons of equivalent noise charge. The noise levels
vary in range of ± 1 ADU for different pixels. The comparable values were achieved also
for PXD5 generation sensors.
Figure 61 – Noise Histogram of Pedestals
6.3
DEPFET Linearity
The relation between drain current Id and charge in the internal gate q is linear
according to Equation 25, when MOSFET in the DEPFET pixel is operated in
saturation mode. Measurement of linearity of the sensor requires well described
adjustment of laser power. Dependence of an optical laser power on laser diode
injection current is not linear. The optical power adjustment by injection current
variation would require precise calibration and would be rather complicated. In the
Mini-matrix system, the laser is used in pulsed operation mode where short laser
pulses are applied to the DEPFET sensor during each integration cycle. Adjustment of
the optical laser power can be easily done by variation of the laser pulse width. When
Page 81/116
The DEPFET Particle Detector Systems and Their Characterization
we assume, that the temperature of the laser module is constant (mean input power is
in the level of µW) than output optical power is proportional to the laser pulse width.
Result of the measurement is shown in Figure 62, where linearity of the DEPFET is
Signal current [nA]
proven.
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
PXD6-450µm
Linear fit
0
200
400
600
800
1000
1200
1400
Laser pulse width [ns]
Figure 62 – Seed response vs. Laser Pulse Width
6.4
Calibration
Detector calibration was done by x-ray sources with discrete energy peaks. Table 1
shows energy of photons, charge deposited in the detector and the branching factor of
the radioactive sources used for calibration. Photons with energies lower than 100 keV
are absorbed in silicon via photo effect as described in Chapter 2.5.2. Their complete
energy is converted to generation of electron-hole pairs. Energy necessary for
generation of one electron-hole pair is Ee-h = 3.65 eV. The number of generated
electron-hole pairs n can be calculated as:
n=
Eγ
Ee − h
,
Equation 31
where Eγ is energy of absorbed photon.
Page 82/116
The DEPFET Particle Detector Systems and Their Characterization
Table 1 – Photon Energies of Radioactive Sources [ 44 ]
Element
Energy
[keV]
Deposited
Branching
Charge
Factor
[electrons]
[%]
55Fe
5.9
1639
109Cd
22
6111
85
25
6944
14
88
24444
4
Figure 63 – Spectra of the Radioactive Sources
Figure 63 presents spectra obtained by irradiation of the PXD5 generation detector
with a
55Fe
source (5.9 keV) and
109Cd
source (22 keV), measured at the Mini-matrix
system in the MPI Munich. Internal amplification of the DEPFET pixels, gq
(Equation 9), can be calculated when ADU is calibrated in the scale of pA. The
calculated internal amplification is in range 368 - 457 pA/e-. Comparable results can be
found in [ 5 ], [ 17 ] and [ 39 ]. The PXD6 generation detector was not calibrated with
the radioactive sources at the Mini-matrix system due to the layout with 2 x 24 active
pixels. All the pixels are at the edges of the active sensor and all hits would be made of
incomplete clusters. The measurement would be distorted. The simple assumption was
done with the laser beam. The same power setting of the red (660 nm) laser gave 20 %
less seed signal for the PXD6 detector than for PXD5 when the laser beam was pointed
in the centre of the pixel. It can be assumed than, that also internal amplification gq is
Page 83/116
The DEPFET Particle Detector Systems and Their Characterization
20 % lower for PXD6, approximately 340 pA/e-. Such assumption is very rough,
because of different layout of the sensors and high sensitivity for position and focus of
the laser.
6.5
DEPFET Gated Operation
Gating the DEPFET is a unique function of the detector which allows making sensor
insensitive for incoming radiation for defined time interval. The charge previously
stored in the internal gate is saved and integration can continue afterwards. This
operation stage is achieved by applying Gate_High (OFF) and Clear_High (ON)
voltages in the same time. In contrast to the real clear, when Gate_Low (ON) voltage is
applied during the clearing and electrons can escape from the internal gate by
thermoionic emission, during the suppressed clear a potential barrier for electrons in
the internal gate is formed and electrons cannot be cleared (Figure 64). The potential of
the internal gate is shifted by a capacitive coupling to the external gate electrode.
The Clear_High (ON) voltage applied during the insensitive (blind) period creates a
shielding potential which deflects trajectories of newly generated electrons and they are
extracted to the clear electrode. Figure 65 shows simulation of electrons trajectories
during the blind mode operation made by R. Richter [ 45 ] (MPI Munich).
Figure 64 – Selectivity of the Clear Process [ 17 ]
Page 84/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 65 – Simulation of Electron Trajectories - R. Richter [ 45 ]
6.5.1
Charge Loss Measurement
The gated mode operation introduces additional clear pulse to the steering sequence.
This pulse might cause additional loses of charge in the internal gate. Therefore,
measurement which determines charge losses due to the additional shielding clear
pulse was carried out. In this test, charge was generated during the integration time,
before the shielding clear pulse. Difference of charge stored in the internal gate with
and without the shielding clear pulse versus generated charge was measured in Figure
66. Figure 67 shows charge loss for variance of clear pulse width. The measurement
was performed with the red (660 nm) laser. Number of electrons lost due to the
shielding clear pulse increases with charge in the internal gate and saturates around
200 electrons. The length of the clear pulse in the insensitive mode was set to 4.8 µs as
a nominal value. Impact of the clear pulse width to charge loss was measured, but no
significant effect was observed within the error of the measurement.
Page 85/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 66 – Charge Loss Due to the Shielding Clear Pulse vs. Charge in the Internal Gate
Figure 67 – Charge Loss Due to the Shielding Clear Pulse vs. Clear Pulse Width
(Generated 12 000 electrons)
Charge loss from the internal gate is affected by setting of the operational voltages. The
critical voltage is Gate_High voltage, which couples electrons in the internal gate
during the blind mode operation. Figure 68 shows simulation of potential in double
pixels structure, where the shift of the internal gate due to the capacitive coupling to
the external gate is visible, when Gate_High (OFF) voltage is applied. When the
external gate voltage is set more negative, internal gate is moving deeper under the
surface of the external gate and the capacitive coupling becoming weaker and for
voltages lower than 3.5 V electrons start loosing from internal gate as visible in Figure
69. Charge loss dependence due to the additional shielding clear pulse as well as
Page 86/116
The DEPFET Particle Detector Systems and Their Characterization
dependence on Back, Clear Gate and Clear_High voltages was also measured, but no
tendencies within the error of measurement were observed.
Figure 68 – Simulation of Potential in the Double Pixel; Left: Gate_ON, Right: Gate_OFF
Figure 69 – Charge Loss Due to the Shielding Clear Pulse vs. Gate_High Voltage
6.5.2 Junk Charge Selectivity
When the DEPFET is switched into the insensitive (blind) mode, charge generated
during this period (junk charge) goes directly to the clear electrode. Small fraction of
the generated charge gets anyway into the internal gate. This effect is called junk charge
selectivity. For measurement of junk charge selectivity, charge was generated by laser
Page 87/116
The DEPFET Particle Detector Systems and Their Characterization
only during the insensitive mode. Figure 70, left displays surface scan with the infrared
laser (1060 nm) of the 450 µm thick matrix. The infrared laser beam penetrates
through full thickness of the sensor, hence, sharp increase of junk charge selection is
observed when the beam hits the internal gate and part of the charge is generated in the
internal gate. Six red points corresponds to six pixel’s internal gates. Figure 70, rights
displays zoom of one pixel. The junk charge selection is affected by settings of
operational voltages and some of them have important role in optimization.
Clear_High voltage, which creates shielding potential, decreases junk charge selectivity
when is set more positive (see Figure 71). This effect can be explained by higher
electrical fields in the bulk structure and higher drift of junk electrons to the clear
electrode. Higher Clear_High voltage also increases clear efficiency, as discussed in
Chapter 6.1.2, but clear voltage cannot be chosen too positive due to the system issues,
which will be described in Chapter 6.5.4.
Another voltage, which influences the junk charge selectivity, is the back voltage (see
Figure 72). Setting with more positive values decreases selection of junk charge.
Explanation can be found in lower reach of shielding clear potential as well as lower
sensitivity of depleted bulk for charge generation.
Figure 70 – Surface IR Laser Scan of the PXD6 450 µm Thick Detector
LEFT: 6 Pixels, RIGHT: 1 Pixel
Page 88/116
The DEPFET Particle Detector Systems and Their Characterization
Junk charge selection versus clear gate voltage is shown in Figure 73. There was
observed strong affection of junk charge selection by clear gate voltage, when clear gate
voltage was set close to the positive values. But it has to be kept in mind that n-channel
starts forming in the clear p-well for positive clear gate voltages and low junk charge
selection is mainly result of the charge loss from the internal gate.
Figure 71 – Junk Charge Selectivity vs. Clear_High Voltage
Figure 72 – Junk Charge Selectivity vs. Back Voltage
Page 89/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 73 – Junk Charge Selectivity vs. Clear Gate Voltage
Figure 74 compares surface red laser scans maxima (seeds) of junk charge selection and
standard signal selection. The junk charge selection maxima are approximately 10 µm
shifted inwards in compare with centres of seeds for normal operation. This is caused
by low junk charge collection in the drift regions of the matrix, which shifts seeds
centres closer to the internal gate. During the standard operation the charge collection
is homogeneous over whole pixel are, hence seed centre is in the middle of the pixel.
This observation leads to conclusion, that for the worst case scenario, single point
measurements should be done over the internal gate location.
Figure 74 – Surface Red Laser Scan of the PXD6 450 µm Thick Detector
In the worst case scenario, incoming particle goes through the area of the internal gate,
leaving fraction of the generated junk charge in the internal gate. Figure 75 shows
Page 90/116
The DEPFET Particle Detector Systems and Their Characterization
results of a single point measurement, when the infrared laser points directly to the
internal gate for different sensor thicknesses and operational voltages. Such
measurement imitates the worst case scenario when the most sensitive part of the
detector for junk charge selection is evaluated. Detector voltages used for the
measurement are in Table 3. The settings called “Optimal” give low junk charge
selectivity and the signal selection is not deteriorated, but the final system limitation is
not taken into account. A fraction of captured electrons of the junk charge generation is
more less constant with number of generated electrons. Fractions of captured junk
charge were measured in Table 2.
Table 2 – Junk Charge Selection
Junk Charge
Thickness
Pixel Size
Voltage
[µm]
[µm x µm]
Settings
PXD6
450
50 x 50
Nominal
3.6
PXD6
450
50 x 50
Optimal
1.6
PXD6
50
50 x 50
Nominal
4.3
PXD6
50
50 x 50
Optimal
3.5
Detector
Selection
[%]
Lower junk charge selection of the thick detector is given by the geometrical
distribution of the detector volume, which is shielded against the junk charge and the
volume near to the internal gate, where generated junk charge is captured. The layout
of thick and thinned detectors is identical, as well as the depth of the internal gate.
Hence, the volume sensitive for the junk charge is similar for the thick, as well as for
the thinned detector. On the other hand, the thinned detector has much thinner bulk
region, where shielding against junk charge is effective. Average junk charge selection
over whole area of the sensor is calculated in the following chapter.
Page 91/116
The DEPFET Particle Detector Systems and Their Characterization
Figure 75 – Single Point Junk Charge Selection vs. Generated Junk Charge
Table 3 – Voltage Settings
6.5.3 Average Junk Charge Selection
The values of junk charge selection calculated in the previous chapter show the worst
case, when the infrared laser beam (or incoming particle) hits the internal gate, where
the sensor is the most sensitive. This region is rather small in compare with overall
pixel area. Figure 76, left demonstrates spatial junk charge sensitivity of the PXD6
Page 92/116
The DEPFET Particle Detector Systems and Their Characterization
thinned matrix with nominal voltage settings. Figure 76, right than highlights regions,
where junk charge selectivity is higher than 2 % of generated junk charge. Estimation of
average junk charge selection can be done, when the pixel area (50 µm x 50 µm) is
divided into two regions, where junk charge selection is lower than 2 % and the region,
where it is more than 2 %. The sensitive region makes approximately 8 % of overall
pixel area. When we assume that there is 4.3 % junk charge selection in the sensitive
area and 0 % over the rest of the pixel the average junk charge selection over the pixel
area is 0.34 %.
Figure 76 – Surface IR Laser Scan of the PXD6 50 µm Thick Detector
LEFT: Seed Response, RIGHT: Junk Charge Selectivity > 2 %
The estimation can be done also for the bigger pixels used in the vertex detector. The
bigger pixels (50 µm x 70 µm) have only bigger drift regions. MOSFET structure as well
as the internal gate stays the same, so we can assume that the sensitive area will be the
same and the ratio of sensitive over non-sensitive area will decrease.
Page 93/116
The DEPFET Particle Detector Systems and Their Characterization
Table 4 – Average Junk Charge Selection (* Extrapolated values)
Average Junk
Thickness
Pixel Size
[µm]
[µm x µm]
PXD6
450
50 x 50
0.29
PXD6
50
50 x 50
0.34
PXD6*
50
50 x 70
0.28
Detector
Charge Selection
[%]
6.5.4 Discussion of the Gated Mode
The results of the gated mode operation presented in the previous chapters
demonstrated possibility of such operation in the view of sensor itself. There still
remain some issues related to the synchronization and system ASICs. The gated
operation has to be synchronized with the accelerator and DCD chips. The DEPFET
sensor can be successfully blinded only, when the blind period comes during the
integration phase of the sensor. During the readout and clearing it is not possible. Due
to the “rolling shutter” readout scheme, there is always part of the detector in readout
phase. It might lead to local inefficiencies. Another option is to pause the readout
during the blind period, which would get the readout of the DEPFET out of synch with
the accelerator.
The gated mode expects to switch on all clear channels at once to start the insensitive
operation. It will lead to high inrush current in the switcher ASICs and will require
large bypass capacitors. High inrush currents might be eliminated by switching on/off
clear channels one by one.
Page 94/116
The DEPFET Particle Detector Systems and Their Characterization
7
CONCLUSIONS
New type of the pixel detector DEPFET is being developed for the Belle II detector at
electron-positron SuperKEKB collider in Japan and future International Linear
Collider. The DEPFET sensor introduces new concept of active pixels with low noise at
room temperature, non-destructive repetitive readout and thinned technologies. The
pixel detector for Belle II consists of two cylindrical layers arranger around the beam
pipe.
The DEPFET test system called Mini-matrix system was designed for testing and
characterization of small DEPFET prototypes. The system has overall noise around 20
electrons of equivalent noise charge and repeatability of the measurement is within the
same error. It can drive sensors with 48 active pixels and the steering sequences are
reconfigurable with high time resolution of 7.5 ns. The test system integrates a
computer controlled positioning of a laser beam focuser which can scan the detector
surface with high resolution of 1.25 µm and computer controlled power supplies, so
automated tests can be performed. The system is placed in the thermally stabilized test
chamber with stability of 0.1 °C, which guarantees precise measurements. The achieved
parameters of the system and test techniques has lower noise, higher spatial resolution
and precision than systems based on ASICs primarily designed for final application
described in [ 5 ], [ 6 ], [ 19 ] and [ 22 ]. The measurements of the DEPFET sensor
performed on the Mini-matrix test system leaded to unique results and deeper
understanding of the DEPFET and were published in refereed journals. Several
versions of the test system were designed and some pieces of the test systems were
supplied to other laboratories.
The DEPFET samples tested on the system had various layouts and thicknesses
(450 µm and 50 µm) as the development of the pixel detector was heading to the final
design. Two generations of the DEPFET sensor PXD5 and PXD6 were tested on the
Mini-matrix test system and the results were used for characterization of the sensor
and new detector designs.
Page 95/116
The DEPFET Particle Detector Systems and Their Characterization
Electrical characteristics of the DEPFET sensor were investigated using the Minimatrix test system. Optimal operational point of the MOSFET cell was found and the
differential output resistance of the DEPFET pixel cell rout = 1.42 MΩ and
transconductance gm = -26.4 µA/V were determined. Clear performance of the
DEPFET sensor was investigated and it was found, that Clear_HIGH voltage has to be
higher than 15 V to achieve complete clear of the internal gate for 450 µm thick sensors.
Sensors with thickness 50 µm required only 12 V for complete clear. Higher values are
recommended. Combination of Clear_LOW and Common Clear Gate voltage helps to
reduce clear voltage swing, but parasitic effects can occur when wrong combination of
these voltages is set. DEPFET sensor works in full depletion so appropriate back
depletion voltage has to be supplied to achieve correct operation. For 450 µm thick
sensors the depletion voltage has to be higher than 120 V. For 50 µm thick sensors the
full depletion was observed at voltages higher than 21 V. Comparable results can be
found in [ 5 ] and [ 6 ], but also new PXD6 DEPFET prototypes wave been investigated
in this thesis.
The edges of the sensor introduce risk of charge loss into the bulk of the sensor or back
injection of charge from balk. The studies of the bulk voltage were carried out and the
optimal bulk voltage 2V was found.
The Belle II design PXD6 matrices have new drift regions added to the enlarged pixels.
It improves drift of electrons from the more remote area of the pixel. The voltage
applied to the drift regions has strong effect on the charge collection of the detector.
The in-pixel studies on charge collection were done and it was found that too positive
values of the drift voltage leads to charge lost from more remote areas of the pixels.
Optimal drift voltage -2 V was found. These results have not been published yet.
Linearity of the DEPFET sensor was proven with pulse width modulated laser and the
sensor was calibrated with radioactive sources. Internal amplification of PXD5
generation matrix was measured in range 368 - 457 pA/e-. Internal amplification of
PXD6 generation matrix was derived from PXD6 by using laser beam as approximately
Page 96/116
The DEPFET Particle Detector Systems and Their Characterization
340 pA/e-. Comparable results achieved with radioactive sources have been presented
in [ 39 ]. Suitability of the pulsed laser beam for precise detectors tests has been shown.
DEPFET gated operation was tested on the Mini-matrix system. This operation allows
making sensor insensitive for incoming radiation for defined time interval. The charge
previously stored in the internal gate is saved and integration can continue afterwards.
This operation stage is achieved by applying positive voltage at the MOSFET gate
electrode and positive voltage at the clear contact in the same time. The positive clear
voltage applied during the insensitive period creates a shielding potential which
deflects trajectories of newly generated electrons and they are extracted to the clear
electrode. Such fast mechanism which can define a time window, where detector stops
integration of new charge, can be used for example to select out noisy bunches injected
in an accelerator. To prove this concept of operation, measurements with red and infra
red laser were carried out on the DEPFET Mini-matrix system. It was proven, that
DEPFET can operate in this way. The average charge selection in the insensitive mode
is lower than 0.4% and the suppressed clear mechanism does not cause charge loss
higher than 200 electrons. Flexibility of the Mini-matrix system, precision and long
term stability leaded to the results which have not been published yet and could not
been obtained on other previous test system based on ASICs.
Development of extended version of the Mini-matrix system already started and it is
expected, that the new version will offer even more detailed measurements of the
DEPFET sensors. DEPFET Collaboration is leading to the end of the development of
the active pixel detector and the expected start date is in 2015.
7.1
Author’s Contribution and Original Results
The presented thesis and its results were obtained in the frame of the DEPFET and in
close touch with other collaborators. This chapter will clearly and briefly point out
author’s contribution and his original results.
Page 97/116
The DEPFET Particle Detector Systems and Their Characterization
The author’s main contribution in the development of the DEPFET pixel detector for
Belle II experiment was design and construction of a low-noise high-precision test
system (Mini-matrix System) for testing and characterization of the DEPFET
prototypes and development of the test techniques. The test system concept was
proposed by the author and the analog front-end readout and precise steering
electronics was developed by the author of this thesis as well as the slow control system.
The PFGA X-Board V2 was previously designed by the Max-Planck Institute in Munich
(MPI) and the sequencer firmware was developed by A. Wassatsch (MPI) according to
the authors requirements. Data acquisition and analysis software were developed by C.
Oswald and P. Kodyš from Charles University in Prague. The main benefits of the
developed Mini-matrix system are low noise of 20 electrons of equivalent noise charge
(ENC) and high precision which allows unique measurements used for characterization
and qualification of the new prototypes of the DEPFET.
All results measured at Mini-matrix system reported in this thesis are author’s original
results. The measurements helped to prove behavior of the DEPFET sensor and its
characteristics as well as investigate new properties. Especially, tests of charge drift in
the PXD6 DEPFET pixels and gated mode operation of the DEPFET sensor brought
original results which were used by the DEPFET collaboration.
Page 98/116
The DEPFET Particle Detector Systems and Their Characterization
8
REFERENCES
[ 1 ] KEMMER, J., LUTZ, G.: New Detector Concepts, Nuclear Instruments and
Methods in Physics Research A253 365-377, 1987
[ 2 ] DOLEZAL, Z., UNO, S., et al.: Belle II Technical Design Report, KEK Report
2010-1, 2010
[ 3 ] The ILD Concept Group: The International Large Detector – Letter of Intent,
KEK Report 2009-6, 2009
[ 4 ] DEPFET Collaboration Home Page:
http://twiki.hll.mpg.de/twiki/bin/view/DEPFET/WebHome
[ 5 ] TRIMPL, M.: Design of a Current Based Readout Chip and Development of
DEPFET Pixel Prototype System for the ILC Vertex Detector, PhD Thesis, Bonn
University, 2005
[ 6 ] KOCH, M.: Development of a Test Environment for the Characterization of the
Current Digitizer Chip DCD2 and the DEPFET Pixel System for the Belle II
Experiment at SuperKEKB, PhD Thesis, Bonn University, 2011
[ 7 ] FISCHER, P., PERC, I., GIESEN, F., KREIDL, C.: Switcher 3 Reference Manual,
Mannheim University, 2007
[ 8 ] Belle II Experiment Home Page: http://belle2.kek.jp
[ 9 ] Belle Experiment Home Page: http://belle.kek.jp
[ 10 ] FLANAGAN, J. W., OHNISHI, Y. et al.: Letter of Intent for KEK Super B
Factory - Part III: Accelerator Design, KEK, 2004
[ 11 ] GLASHOW, S. L.: Partial Symmetries of Weak Interactions, Nucl. Phys. 22 579,
1961
[ 12 ] WEINBERG, S.: A Model of Leptons, Phys. Rev. Lett. 19 1264, 1967
[ 13 ] SALAM A.: Weak and Electromagnetic Interactions, In the Proceedings of 8th
Nobel Symposium, Lerum, Sweden, pp 367-377, 1968
Page 99/116
The DEPFET Particle Detector Systems and Their Characterization
[ 14 ] Home Page of the LHC: http://lhc.web.cern.ch/lhc
[ 15 ] Home Page of the ATLAS Experiment: http://www.atlas.ch
[ 16 ] Home Page of the CMS Experiment: http://cms.web.cern.ch
[ 17 ] DOLEZAL, Z., KIESLING, C., LACASTA, C., MOSER H.-G. et al.: The Belle II
PXD Whitebook, DEPFET Collaboration, 2012
[ 18 ] PERIC, I. et al.: DCDb and Switcherb, the Readout ASICs for Belle II DEPFET
Pixel Detector, IEEE Nucl. Sci. Symposium Conference Record N31-5, pp. 15361539, 2011
[ 19 ] PERIC, I. et al.: DCD - the Multi-channel Current-mode ADC Chip for the
Readout of DEPFET Pixel Detectors, IEEE Trans, Nucl. Sci. vol. 57, no. 2, pp.
743-753, 2010.
[ 20 ] OSWALD, C.,: Tests of DEPFET pixel detectors and Simulation of the Belle II
vertex detector, Diploma Thesis, Bonn University, 2011
[ 21 ] MOSER H.-G.: Silicon detector systems in high energy physics, Progress in
Particle and Nuclear Physics, 63(1):186–237, 2009.
[ 22 ] NICULAE, A. S.: Development of a Low Noise Analog Readout for a DEPFET
Pixel Detector, PhD Thesis, Siegen University, 2003
[ 23 ] ANDRICEK, L., FISCHER, P., HEINZINGER, K., et. al.: The MOS-Type
DEPFET Pixel Sensor for the ILC Environment, Nuclear Instruments and
Methods in Physics Research. Elsevier Science, 2003
[ 24 ] GÄRTNER, K., RICHTER, R.: DEPFET Sensor Design Using an Experimental
3d Device Simulator, Elsevier Science, 2006
[ 25 ] SANDOW, C., ANDRICEK, L., FISCHER, P. et. al.: Clear-performance of Linear
DEPFET Devices, Elsevier Science, 2006
[ 26 ] CHU, J. L.: Thermionic Injection and Space-charge-limited Current in Reachthrough p+np+structures, Journal of Applied Physics, 1972
Page 100/116
The DEPFET Particle Detector Systems and Their Characterization
[ 27 ] RICHTER, R., et al.: Design and Technology of DEPFET Pixel Sensors for
Linear Collider Applications, Nucl. Instrum. Methods Phys. Res., Sect. A,
511:250–256, 2003
[ 28 ] Octopus CompuScope 83XX - 14-Bit Family of Multi-channel Digitizers for the
PCI Bus Datasheet, www.gage-applied.com, 2006
[ 29 ] HAELKER, O.: X-Board V2 Documentation, MPI Munich, 2007
[ 30 ] WASSATSCH, A.: Mini-Matrix-Sequencer Manual Version 1.0, MPI Munich,
2009
[ 31 ] SCHEIRICH, J., OSWALD, C., KODYS, P.: First Measurement on the DEPFET
Mini-Matrix Particle Detector System, Conference Proceedings of the Eighth
International
Conference
on
Advanced
Semiconductor
Devices
and
Microsystems, IEEE Bratislava, 2010
[ 32 ] SCHEIRICH, J.: Design of Readout Electronics for the Mini-matrix DEPFET
Detector, Master’s Thesis, Czech Technical University in Prague, 2008.
[ 33 ] Home Page of the MEDIPIX: http://medipix.web.cern.ch/medipix/
[ 34 ] URBÁŘ, J., SCHEIRICH, J., JAKUBEK, J.: Medipix/Timepix cosmic ray tracking on
BEXUS stratospheric balloon flights, Nuclear Instruments and Methods in Physics
Research A, 2011
[ 35 ] WHITE, W. H., LAMPE, D. R., BLAHA, F. C., MACK, I. A. Characterization of
surface channel CCD image arrays at low light levels, IEEE Journal of SolidState Circuits Vol.SC-9, pp.1-14, 1974
[ 36 ] BUTTLER, W., HOSTICKA, B.J., LUTZ, G.: Noise Filtering for Readout
Electronics, Nuclear Instruments and Methods in Physics Research A288 187190, 1990
[ 37 ] MIX, D. F., SCHMITT, N.M.: Circuit Analysis for Engineers, Continuous and
Discrete Time Systems, John Wiley & Sons, New York, 1985
Page 101/116
The DEPFET Particle Detector Systems and Their Characterization
[ 38 ] ANDRICEK, L. et al.: Intrinsic Resolutions of DEPFET Detector Prototypes
Measured at Beam Tests, Nuclear Instruments and Methods in Physics
Research A 638 24–32, 2011
[ 39 ] RUMMEL, S., ANDRICEK, L., MOSER, H.-G., RICHTER R. et al.: Intrinsic
Properties of DEPFET Active Pixel Sensors, Journal of Instrumentation, 2008
[ 40 ] TSIVIDIS, Y.: Operation and Modeling of the MOS Transistor, Oxford
University Press, 1999
[ 41 ] LUTZ, G.: Semiconductor Radiation Detectors, Springer Verlag, Berlin, 1999
[ 42 ] CHAN, T.Y., CHEN, J., KO, P.K., HU, C.: The Impact of Gate-Induced Drain
Leakage Current on MOSFET Scaling, University of California, Berkeley, IEEE,
1987
[ 43 ] MOLL, J.L.: Physics of Semiconductor, McGraw Hill, P. 253, 1964
[ 44 ] KOHRS, R.: Development and Characterization of a DEPFET Pixel Prototype
System for the ILC Vertex Detector, PhD thesis, University Siegen, 2008.
[ 45 ] RICHTER, R.: DEPFET Simulations, 2012
Page 102/116
The DEPFET Particle Detector Systems and Their Characterization
APPENDIX A LIST OF CANDIDATE’S WORKS
RELATING TO THE DOCTORAL THESIS
Impacted Publications on Web of Science
[A1] ANDRICEK, L., KODYŠ, P., KOFFMANE, C., NINKOVIC, J., OSWALD, C., RICHTER, R.,
RITTER, A., RUMMEL, S., SCHEIRICH, J.*, WASSATSCH, A.: Advanced Testing of the
DEPFET Minimatrix Particle Detector, Journal of Instrumentation, 2012
[A2]ANDRICEK, L., CARIDE, J., DOLEŽAL, Z., DRÁSAL, Z., ESCH, S., FREY, A.,
FURLETOVA, J., FURLETOV, S., GEISLER, C., HEINDL, S., IGLESIAS, C.,
KNOPF, J., KOCH, M., KODYŠ, P.*, KOFFMANE, C., KREIDL, C., KRÜGER, H.,
KVASNIČKA, P., LACASTA, C., MALINA, L., MARIÑAS, C., NINKOVIC, J.,
REUEN, L., RICHTER, R. H., RUMMEL, S., SCHEIRICH, J., SCHNEIDER, J.,
SCHWENKER, B., VÁZQUEZ, P., VOS, M., WEILER, T., WERMES, N.: Intrinsic
resolutions of DEPFET detector prototypes measured at beam tests, Nuclear
Instruments and Methods in Physics Research A, 2011
[A3]URBÁŘ, J.*, SCHEIRICH, J., JAKUBEK, J.: Medipix/Timepix cosmic ray
tracking on BEXUS stratospheric balloon flights, Nuclear Instruments and
Methods in Physics Research A, 2011
Refereed Scientific Journals
[A4] SCHEIRICH, J.*: The DEPFET Mini-matrix Particle Detector, Acta Polytechnica,
ČVUT Press, Praha, 2010
_____
*
Corresponding author
Page 103/116
The DEPFET Particle Detector Systems and Their Characterization
Other Publications
[A5] SCHEIRICH, J.*, OSWALD, C., KODYŠ, P.: First measurement on the DEPFET Minimatrix particle detector system, ASDAM, IEEE, 2010
[A6] URBÁŘ J.*, SCHEIRICH J., JAKŮBEK J.: MEDIPIX Cosmic Ray Tracking Device on
BEXUS-7 Stratospheric Balloon Flight, Proceeding of the 19th ESA Symposium on
European Rocket and Ballon Programes and Related Research, European Space Agency,
Noordwijk NL, 2009
[A7] SCHEIRICH, J.*: DEPFET Mini-matrix Particle Detektor, POSTER 2010 - Proceedings
of the 14th International Conference on Electrical Engineering, CTU in Prague, FEE, 2010
[A8] VOS, M.*, FUSTER, J., LACASTA, C., MARINAS, C., DIEGUEZ, A., GARRIDO, L.,
GASCON, D., COMERMA, A., FREIXES, L., CASANOVA, R., VILELLA, E., RIERABABUES, J., VILASIS-CARDONA, X., GASPAR, A., PAJARES, J., RODRIGUEZ, P.,
FURLETOVA, J., FURLETOV, S., KOHRS, R., KOCH, M., KRÜGER, H., REUEN, L.,
SCHNEIDER, J., WERMES, N., FISCHER, P., KREIDL, C., PERIC, I., KNOPF, J.,
LANGE, S., KÜHN, W., MÜNCHOW, D., FREY, A., GEISLER, C., SCHWENKER, B., DE
BOER, W., BARVICH, T., BROVCHENKO, O., HEINDL, S., SIMONIS, H.J., WEILER, T.,
BRODZICKA, J., BOZEK, A., KAPUSTA, P., PALKA, H., ANDRICEK, L., CHEKELIAN, V.,
KIESLING, C., KOFFMANE, C., LU, S., LUTZ, G., MOLL, A., MOSER, H.G.,
NEDELKOVSKA, E., NINKOVIC, J., PROTHMAN, K., RICHTER, R., RITTER, M.,
RUMMEL, S., SIMON, F., DOLEZAL, Z., DRASAL, Z., KODYS, P., KVASNICKA, P.,
SCHEIRICH, J., CARIDE, J., ESPERANTE, D., GALLAS, A., PEREZ, E., RODRIGUEZ, P.
VAZQUEZ, P.: DEPFET active pixel detectors, Proceedings of Science, 2009
_____
*
Corresponding author
Page 104/116
The DEPFET Particle Detector Systems and Their Characterization
Technical Reports
[A9] DOLEŽAL, Z., KIESLING, C., LACASTA, C., MOSER H.-G. et al.: The DEPFET
PXD Whitebook, DEPFET Collaboration, 2012
[A10] DOLEŽAL, Z., UNO, S., et al.: Belle II Technical Design Report, KEK Report
2010-1, 2010
[A11] The ILD Concept Group: The International Large Detector – Letter of Intent,
KEK Report 2009-6, 2009
In Preparation
[A12] ANDRICEK, L., DOLEŽAL, Z., KOFFMANE, C., KODYŠ, P., MOSER, H-G.,
MÜLLER, F., NINKOVIC, J., OSWALD, C., RICHTER, R., RITTER, A.,
RUMMEL, S., SCHEIRICH, J.*, WASSATSCH, A.: Laser Tests of the DEPFET
Gated Operation, Journal of Instrumentation, 2013 (expected)
[A13] RICHTER, R.*, ANDRICEK, L., BÄHR, A., GÄRTNER, K., KIESLING, C.,
KOFFMANE, C., KRÜGER, H., KREIDL, C., MOSER, H.-G., MÜLLER, F.,
NINKOVIC, J., PERIC, I., RUMMEL, S., SCHEIRICH, J., SCHWENKER, B.,
WILK, F.: Particle Detection with DEPFET Arrays in Gated Mode, IEEE
Nuclear Science Symposium and Medical Imaging Conference, 2013 (expected)
_____
*
Corresponding author
Page 105/116
The DEPFET Particle Detector Systems and Their Characterization
APPENDIX B RÉSUMÉ
Jan Scheirich was born in 1983 in Poprad, Slovakia. He spent his bachelor, master
and doctoral studies at the Czech Technical University in Prague. His doctoral studies
were carried out in cooperation with Charles University in Prague where he
participated on development of DEPFET particle detector for Belle II experiment in
Tsukuba, Japan. He was selected to CERN Summer School Programme in 2007 where
he developed fast high voltage switches for RFQCB (Radio Frequency Quadrupole ion
Cooler and Buncher) at ISOLDE experiment. He participated in two successful students
stratospheric balloon campaigns BEXUS-7 and -9 (Balloon EXperiments for University
Students) carrying MEDIPIX and TIMEPIX particle detectors, sponsored by German
Aerospace Center, Swedish National Space Board and European Space Agency in 2008
and 2009. The launch campaigns took place in Swedish Space Center ESRANGE. He
was offered a research fellowship by the European Space Agency in 2010, where he did
research on a Power Distribution Unit for the Large Space Simulator situated in the
European Space Agency's European Space Research and Technology Center in the
Netherlands. His PhD research was partially financed by three grants (2010 - 2012)
which he had won in the Student Grant Competition of the Czech Technical University
in
Prague
(GS10/075/OHK3/1T/13,
SGS11/066/OHK3/1T/13,
GS12/073/OHK3/1T/13). He participated on a research grant of Czech Science
Foundation (203/10/0777 2010 – 2012) and on a grant of Foundation of Ministry of
Education, Youth and Sports of the Czech Republic (LA10033 2010 – 2012). He was
awarded by Program Committee of POSTER 2010 and the MTT/AP/ED/EMC Joint
Chapter of the Czechoslovakia Section IEEE at 14th International Student Conference
on Electrical Engineering POSTER 2010 for the work ”DEPFET Mini-matrix Particle
Detector”. He is member of DEPFET Collaboration and he has been collaborating with
Max-Planck Institute in Munich, Bonn University, Ludvig-Maximilians University in
Munich, KEK Laboratory in Japan, CERN in Switzerland and others during his PhD
studies.
Page 106/116
The DEPFET Particle Detector Systems and Their Characterization
APPENDIX C MINI-MATRIX V2 DOCUMENTATION
Figure A1 – Mini-matrix Setup
JTAG:
X-Board firmware upload
X-Board 5V:
Digital voltage
Trigger Volt. Adj.: Adjustment of the output trigger voltage level (5V Max.) Measure
at “TP 2 3.3V” test point.
Triggers:
High impedance output. Do not use at 50 ohm.
High Voltage:
Bias voltage. 1Meg ohm serial resistor is included.
GND J24:
Jumper connecting AGND and DGND
DEPFET
Mini-Matrix:
See fig. A3
Individual Subtr.
Adj.:
Adjustment of subtraction voltage for individual channels.
Amplifiers:
Caution: No lock on the slot. Check PIN1 position
Offset Adj.:
Adjustment of the offset voltage of the second amp stage.
SubtrCOM Adj.:
Adjustment of the internal SubtractionCOM voltage
Analog output
Channels:
50 ohm amplifiers outputs going to ADC card
SubtrCom J29:
Internal power supply configuration jumper.
OPEN = External power supply; CLOSED = Internal power supply
Do not connect external power supply, when jumper is closed.
Power:
Power supply connector (See fig. A2)
Virt. Drain Adj.
Adjustment of virtual drain voltage
Virt. Drain J15:
Internal power supply configuration jumper.
Page 107/116
The DEPFET Particle Detector Systems and Their Characterization
OPEN = External power supply; CLOSED = Internal power supply
Do not connect external power supply, when jumper is closed.
X-Board sequence configuration port. Virtual serial port. Download
a driver at: http://www.ftdichip.com/Drivers/VCP.htm
Check for COM port ID at “Control Panel -> System -> Hardware
-> Devices -> Ports (COM and LPT)”
USB:
P1
Vds
C142
100n
Vclear
C136
10u
C138
100n
C147
10u
C151
100n
C154
10u
Bulk
VclearLOW
Edge
Vgate
C149
100n
AB2
C158
100n
C153
10u
C156
10u
Vv irtdrEXT
VgateLOW
AB1
C157
100n
C159
10u
Ccg
C164
100n
C163
10u
C165
100n
C166
100u
C185
100n
C186
1m
VsubtrCOM
+27VIN
C184
100n
C167
1m
-20VIN
+6VIN
C188
100n
19
37
18
36
17
35
16
34
15
33
14
32
13
31
12
30
11
29
10
28
9
27
8
26
7
25
6
24
5
23
4
22
3
21
2
20
1
C187
1m
CONNECTOR DB37
WRNING: Connector pinout doesn't match
with the previous system MIMA V1
Figure A2 – Power Connector
Page 108/116
The DEPFET Particle Detector Systems and Their Characterization
Figure A3 – DEPFET Connector
Figure A4 – Mima <-> X-Board Connector
Page 109/116
The DEPFET Particle Detector Systems and Their Characterization
Mean
Current
Voltage Current Limit
# Name
[V]
[mA]
[mA]
Note
1 X-Board
5
500
1100 Digital
2 HV-Bias
180
0
0
3 Fan
12
100
250 Fan is optional
4 "+6V"
6
3
10 Digital
5 "-20V"
-18
200
800 800mA is required for start up!
6 "+27V"
11
200
800 800mA is required for start up!
Exact voltage has to be adjusted according to the
7 VsubtrCOM
10,1
80
150 configuration. Internal voltage source can be used.
8 Ccg
-0,9
0
1
9 AB1
N/A
N/A
N/A
Unused
10 AB2
N/A
N/A
N/A
Unused
11 Eddge
N/A
N/A
N/A
Unused
12 VgateLOW
-4
35
50
13 Vgate
6
35
50
14 VclearLOW
3,8
35
50
15 Vclear
16,8
35
50
16 Vsd
-5
1
10 Internal voltage source can be used.
17 Bulk
10
0
1
Initial System Voltages for PXD5 Matrix
Page 110/116
The DEPFET Particle Detector Systems and Their Characterization
APPENDIX D SCHEMATICS
C1
3pF
R1
13k
VDD-
R5
U2
OUT
+
10
7
+
U1
R7
39k
0
C4
EL 21 26 C_ RE B
E L21 26 C_ RE B
7
-
Vout+
OUT
-
4
Input
4
VBB+
VDD+
VBBR10
39k
R12
4.7u
C18
4.7u
C19
Vsubtr
4.7u
C20
2.1k
Vv irtdr
C6
4.7p
Vof f set
J1
L1
1
+15V
2
VDD+
Input
10uH
100nF
C5
Vsubtr
4.7u
C7
Vv irtdr
Vof f set
Vout-
L3
1
2
VBB+
Vout+
10uH
+15V
100nF
C8
4.7u
C9
-15V
+6V
-6V
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
CON20
L5
1
-15V
2
VDD-
10uH
100nF
C11
4.7u
C12
L6
1
2
VBB-
10uH
100nF
C13
4.7u
C10
Amplifier V2
Page 111/116
The DEPFET Particle Detector Systems and Their Characterization
Main Board V2
Page 112/116
The DEPFET Particle Detector Systems and Their Characterization
Slow Control
Page 113/116
The DEPFET Particle Detector Systems and Their Characterization
VsubtrCOM
J54
R47
0R
Jumper3x8
R50
0R
R53
0R
R56
0R
R59
0R
R62
0R
R65
0R
R68
0R
R48
5k
R51
5k
R54
5k
R57
5k
R60
5k
R63
5k
R66
5k
Vsubtr8
Vsubtr7
Vsubtr6
Vsubtr5
Vsubtr4
Vsubtr3
Vsubtr2
Vsubtr1
R69
5k
R49
0R
R52
0R
R55
0R
R58
0R
R61
0R
R64
0R
R67
0R
R70
0R
J55
1
2
AGND
Subtraction Board
Page 114/116
The DEPFET Particle Detector Systems and Their Characterization
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
J1
CH1 A
1
J2
CH1 B
1
RL1
CH2 A
1
5
1
J5
CH2
1
RB751V40T1G
VCC
D1
C1
100n
2
3
J6
CH1
1
8
4
7
J4
J3
6
CH2 B
1
5
4
3
2
5
4
3
2
20-2051-DPDT
J7
CH3 A
1
J8
CH3 B
1
Q1
2N7002
RL2
J9
6
4
7
J10 CH4 A
1
CH3
1
5
4
3
2
5
4
3
2
8
J11 CH4
1
5
1
RB751V40T1G
VCC
D2
C2
100n
2
3
J12 CH4 B
1
20-2051-DPDT
Q2
2N7002
J13
1
2
R1
D3
2
470R
1
J14
STATE B
TTL
LED
J15
1
2
VCC
C3
100n
C4
10uF
Vcc
Multiplexer Board
Page 115/116
The DEPFET Particle Detector Systems and Their Characterization
Main Board V3
Page 116/116
The DEPFET Particle Detector Systems and Their Characterization
Download