The most complex
Silicon Detectors:
Silicon Drift!
Silicon Drift Detectors (SDD)
E. Gatti, P. Rehak, Semiconductor Drift Chamber - An Application of a Novel
Charge Transport Scheme, Nucl. Instr. and Meth. A 225, 1984, pp. 608-614.
I rivelatori a deriva sono usati
– nell'ambito degli esperimenti con ioni pesanti ultrarelativistici (altissima desità di
particelle, tracking con TPC, rivelatori lenti read-out time = ms)
– per applicazioni di tipo medico / industriale (rivelazione raggi X)
P.
Lechner et al., Silicon drift detectors for high resolution room temperature X-ray
spectroscopy, Nucl. Instr. and Meth. 1996; A 377, pp. 346-351.
Vantaggi
+
+
+
+
punti 2D senza ambiguità
alta risoluzione per entrambe le
coordinate
costo minore rispetto ai pixel
lettura analogica Particle ID
Svantaggi / sfide
-
tempo di raccolta lungo
elettronica complessa
alta tensione (per i silici)
richiede calibrazione accurata
della temperatura
richiede calibrazione accurata
dell'uniformità del silicio
Rivelatori al silicio a
deriva
Il concetto
1. due giunzioni p+n contrapposte ("back-to-back"), entrambe in
polarizzazione inversa, svuotano il silicio dalle cariche libere e
creano un campo elettrico con un minimo al centro
d
d
2
2
d
- VD
- VD
4
p+
n
n
p+
4
- VD
4
+
0
=
2. creando una catena di coppie di giunzioni "back-to-back", ciascuna
coppia con una tensione decrescente rispetto a quella precedente, si
genera il campo di deriva per portare la carica fino all'elettrodo n+
-100
-90
y
-100
-90
-80
-70
-60
-50
-80
-40
z
-70
-100
-90
VD = 50 V
d = 300 µm
-80
-70
-60
+
p
-50
+
n
-40
-60
-50
-
n
-50
-50
-40
z
y
-V
Silicon Drift Detectors (1)
Silicon drift detectors are charged partcle detectors
capable of providing both two-dimensional position
information and ionization measurements.
The operating principle is based on the measurement of
the time necessary for the electrons produced by the
ionization of the crossing particle to drift from the
generation point to the collection anodes, by applying an
adequate electrostatic field.
7 cm
The transport of electrons, in a direction parallel to
the surface of the detector and along distances of - VD
several centimetres, is achieved by creating a drift 4 p+
- VD
channel in the middle of the depleted bulk of a
4
silicon wafer. At the edge of the detector, the
electrons are collected by an array of small size
0
anodes.
d2
d2
n
n
+
- VD
4
+
p
d
=
Particle
n+
n+
n+
P+
P+
P+
n
P+
P+
P+
P+
P+
+- +
- +-
P+
P+
x
y
P+
The measured drift time gives
information on the particle
impact point coordinate y. The
charge sharing beween anodes
allows the determination of
the coordinate along the anode
direction x.
Silicon Drift Detectors (2)
In practice, both the drift field and the depletion bias are produced by p+ parallel strips
implanted on both faces of the detector. Each strip is polarized with a negative voltage
proportional to its distance from the anodes, in order to produce the drift field. In this
way, the p+-n junctions are reverse polarized and can assure the depletion of the detector
through a n+ ring placed at the periphery of the detector.
Coordinate x (anode axis)
The diffusion and Coulomb repulsion between
electrons play a significant role in the drift
detectors since the drift time is of the order
of a few mm. In the thickness direction, they
are compensated by the parabolic potential, but
generate an increase of the electron cloud size
in both other directions. The electron cloud
reaches the collection zone with a size
increasing as a function of the total drift time.
Thus a charge may be collected by more than
one anode and the coordinate x is determined
as the centroid of the charge deposited on the
touched anodes. Typically, a 200 mm pitch
allows a precision of 30 mm.
Coordinate y (drift axis)
Potential energy in the Silicon Drift
Detector obtained with a numerical simulation
ANODE
The signal measured on each anode is amplified and sampled with a typical frequency of a few
tens of MHz, depending on the drift velocity and the peaking time of the electronics. The
coordinate y is measured by calculating the elapsed time between an external trigger and the
arrival of the charge
Applicazioni mediche
diagnostica con raggi X
It consists of a volume of fully depleted
high-resistivity silicon, in which an electric
field with a strong component parallel to
the surface drives electrons generated by
the absorption of ionising radiation
towards a small sized collecting anode.
The electric field is generated by a
number of increasingly reverse biased
field strips covering one surface of the
device
The radiation entrance window on the
opposite side is made up by a nonstructured shallow implanted junction
giving a homogeneous sensitivity over the
whole detector area.
animation
•The unique property of this type of detector is the extremely small value of the anode capacitance,
which is independent of the active area. This feature allows to gain higher energy resolution at
shorter shaping times compared to conventional photo diodes and Si(Li) detectors, recommending the
SDD for high count rate applications
• Due to the elaborated process technology used in the SDD fabrication the leakage current level is
so low that the SDD can be operated with moderate cooling by means of a single stage Peltier
element.
•The SDD's energy resolution (FWHM < 145 eV @ MnKa, -20oC) can be compared to that of a Si(Li)
detector requiring no expensive and inconvenient liquid nitrogen cooling. It surpasses the quality of
pin-diodes.
Esperimenti di fisica nucleare delle
alte energie: CERES - STAR
• CERES-NA45 @ CERN SpS
– Sistema di Drift circolari usate come rivelatore di vertice (19962000)
– Problemi legati al partitore esterno
• STAR @ RHIC: Silicon Vertex Telescope
– 3 piani cilindrici di SDD rettangolari con partitore esterno
– smontate dopo 4 anni di presa dati
– Problemi legati al partitore esterno
SDD per ALICE
R&D
• inizio progetto 1992
– INFN DSI project in collaboration with CANBERRA Semiconductors. The
aim of the project was the production of a large area SDD (5 inch wafers)
with integrated high voltage divider
• studio del materiale
– Neutron Transmutation Doped 5-inch silicon wafers with a resistivity of
3kΩ*cm and a thickness of 300μm.
• simulazione del campo di deriva e raccolta
– definizione della geometria di catodi, anodi e metallizzazione
– definizione dei parametri di funzionamento V/gap = tensione di
polarizzazione
• simulazione delle zone di guardia
– raccolta corrente superficiale
– minimizzazione rischi di break-down tra catodi centrali e anello di guardia
• test prototipi e studi di radiation damage
– Primi prototipi 1993-94  rivelatori unidirezionali
– Primo prototipo con geometria quasi definitiva bidirezionale 1998
Simulazioni (I)
• We need an extensive numerical simulation of SDD electrical
behaviour.
• The simulation of a complete large area detector cannot be
performed because of the huge amount of memory required. Hence
the calculation has to be done on limited portions of the device,
introducing artificial boundary conditions.
• We use a 2D approximation ( 3D differential equations would in
principle be required)
– This approximation can be done if we consider cross sections orthogonal to the
cathodes in such a way that the partial derivative along the third dimension is
negligible.
• We use ATLAS device simulation software produced by SILVACO.
The simulations regard the following regions:
–
–
–
–
collection zone
drift region
guard region
injectors
Simulazioni (II-drift region)
• The electric field needed to drift the electron cloud is
imposed by a suitable bias of the drift cathodes
• In order to obtain a realistic result, the lateral
boundaries must be kept far from the region of
interest (the central region).
• For this reason we used a high
number of cathodes (9),
leading to a total length of
about 1.5mm against a device
depth of 0.3mm.
• Moreover, the first couple of
cathodes on the left are
extended in order to further
remove the left edge.
• An anode is located at the
right side in order to
guarantee a contact for the
bulk.
Simulazioni (III-drift region)
•
•
•
The bottom of the potential gutter is
perfectly linear; The green line is the
potential profile 0.1μm under the
Silicon oxide.
It is visible the effect of the fieldplate that lowers the potential
variations (electric field) near the
junctions
It is worthwhile noting that the
distance between the red and the
green line, passing from one cathode
to another, is constant, meaning that
there is no influence of the boundary
solution.
Simulazioni (IV-Collection zone)
• The collection zone is one of the critical regions in a SDD. Here
the electron cloud is forced to move from the middle plane of
the detector toward the anode array in the n-side. The forcing
electric field is applied by properly biasing the last few
cathodes in the proximity of the anodes.
– First we must avoid a trapping
of the signal electrons under
the oxide when approaching
the surface.
– Second we have to minimize
the non-linearity of the drift
speed associated with the
transversal movement towards
the n-side of the detector.
– Third we have to guarantee a
good potential separation
between anodes and perimeter
to avoid inefficiencies of the
electron collection.
Simulazioni (IV-Collection region)
.The picture shows an overlay of the potential map and the trajectory
of eight electrons placed at various positions.
It is worthwhile noting that the "pull-up" region is very short (200μm),
minimising the systematic error on the drift time. Such a kind of
collection minimizes also the risk of electron trapping under the
oxide because the trajectories are kept far from the surface up to
the anode
The potential barrier between anodes and n+bulk contact is about 6V.
Deriva
Deriva
ALICE
DB-2
SDD Alice: 7.02  7.53 cm  300mm
Diviso in due volumi attivi dal catodo centrale polarizzato a -2370V
Ciascun volume ha 256 anodi di raccolta (passo di 294 mm).
292 catodi (passo di 120mm) permettono di impostare il campo di deriva
tramite un partitore di resistenze impiantate.
Caratteristiche principali
partitore integrato
sul rivelatore
doppia catena di
resistenze in
polisilicio
(R=160kΩ)
– una per catodi di
campo
– una per catodi di
guardia
anodi
griglia isolante
Guard Zone
• At each side of the drift cathode array, p+
implants (guard strips) grade the potential from
the highest negative voltages to grounded outer n+
implant ring. Since this region should be as small
as possible the electric field needs a careful
evaluation.
• Furthermore, as the voltage difference between
two consecutive guard-zone strips is 16V (see
detector description), the punch-through
phenomenon should be carefully evaluated
Caratterizzazione dei rivelatori
Produzione iniziata settembre 2004 (Canberra Semiconductors)
Test sui rivelatori nudi svolti presso INFN Trieste
Calculating on a yield of 50%, about 500 detectors had been characterized at best
– First a visual inspection has to be done to check for interruptions or shorts in the metal.
– The second step is to check both the current at the anodes
the probe card connects together the anodes in groups of eight
–
and the linearity of the potentail on the divider
checking the voltage drop every ten drift cathodes
Connection made by use of
two probe cards able to
contact the detector safely
from both faces (probe pad
coordinates) to choose only
the well-performing ones with
minimal risk of damage to the
detector during this operation.
Risultati
tipici
uniformità del partitore
corrente agli anodi
Detector selection criteria
• Linearity of the potential distribution on the integrated divider.
– Due to local defect generating high current or to a punch-through current
among the cathodes.
– Non-linearity of the potential distribution generates a systematic error on the
position resolution along the drift direction. Furthermore when the distortion
on one side
– If the difference on the voltage drop on the resistor connecting two
consecutive cathodes is of the order of 0.1V  the electron cloud is shifted
dangerously to one of the surfaces
• anode leakage current
– this determines the noise - and therefore efficiency - of the detector readout
(the anode capacitance can be ignored since it will always be small compared
to the fixed contribution introduced by the readout microcables).
– LIMIT for Anode current = 100 nA
Esempio di acquisizione di segnali
Segnale (ADC)
Lettura di un canale (impulsi di test)
Tempo  25ns
Canale
Lettura di 64 canali (impulsi di test)
Tempo  25ns
Tempo
deriva
Esempio di un evento
dall'esperimento STAR
Anodo
Ampiezza
Problemi delle SDD
• Sensibilità alla temperatura
– l'integrazione di strutture di calibrazione
• Sensibilità all'uniformita del silicio di base
– la mappatura dei sensori
Temperature stability dependence
La velocità di deriva dipende dal campo, ma anche dalla temperatura
(per non dover correggere per l'influenza della temperatura sulla misura, ci
vorrebbe una stabilità di 0.1 gradi Celsius)
=> per calibrare, strutture chiamate "iniettori" sono integrate sul rivelatore
permettendo di generare elettricamente segnali da posizioni conosciute
667 V/cm
120 ns
3%
T=3.6°
Iniettori
- 3 injector lines are inserted
between
consecutive
drift
cathodes at distances of 3 mm,
17.6 mm and 34 mm from the
anodes.
- Each line consists of a metal
strip deposited over the oxide.
Beneath the strip, separated by
a 100 nm thick oxide, it runs a
p+ implant interrupted in 33
points that constitute the
injection locations.
- In these points, 100 mm wide,
there is an accumulation of
electrons due to the positive
oxide
charge.
Applying
a
negative pulse to the metal line
we push a certain number of
these electrons in the silicon
bulk.
- At the anodes we obtain three
sets the drifted images of the
33 injectors.
iniettori
Bias scheme for MOS charge injector
pulse
generator
Injector event in SDD module mounted on the ladder
simulazione
e risultati dei
test di
laboratorio
Uniformita’ del drogaggio
(resistivita’)
• Se la resistività del silicio del
rivelatore non è abbastanza
uniforme si creano dei campi
elettrici parassiti che spostano
la carica dalla traiettoria
ideale,
• quindi si trova un errore
sistematico fra la posizione
misurata e la posizione dove la
particella è realmente passata
Mappe degli errori sistematici per
metà di un rivelatore
x>0
(beam test data)
x = XSDD-XREF
x<0
y>0
y = vdrift*tSDD-YREF
y<0
Scelta del materiale
• studi accurati su silicio “Floating zone” e “Neutron
Transmutation Doped NTD” 1992-1994
– Floating zone: fluttuazioni di resistività fino a 30%
– NTD: <10% (Silicio Wacker)
• Produzione ALICE:
– iniziata su Wacker -> Fluttuazioni viste in slide precedente 
necessità di “mappare” ogni singolo rivelatore per correggere offline i dati dagli effetti sistematici (stabili nel tempo)
– continuata su TOPSIL  fluttuazioni pressoche’ inesistenti
• stazione di mappatura presso lab tecnologico
INFN Torino usata come stazione di test su
“moduli” completi (DETECTOR + FEE + CAVI
ALIMENTAZIONE E DATI + schede ausiliarie)
SDD: front-end
• FEE amplify-memory-ADC (PASCAL)
Analogue
memory
ADCs
Preamplifiers
• event buffer chip (AMBRA)
SDD module p-side
with ladder cables and end-ladder boards
Transition
Cable
LV Board
HV
Board
Heat exchanger at the back of Hybrid
LV Board
L’Inner Tracking System di
ALICE
The ALICE Inner Tracking System
SSD
SDD
SPD
Lout=97.6 cm
Rout=43.6 cm
• 6 Layers, three technologies (keep occupancy ~constant ~2%)
– Silicon Pixels (0.2 m2, 9.8 Mchannels)
– Silicon Drift (1.3 m2, 133 kchannels)
– Double-sided Strip Strip (4.9 m2, 2.6 Mchannels)
ITS Mechanical assembly
Positioning rings
CF support cones
Reverse biased p-n junction (I)
-V
p
n
+- - +- +- - + -+ - +
-xp
+ + +-++ + +- + + + +-+d
+V
xn
p+
 2V E

  e( N D  N A )
 2 

x
x  r  0
eN
E p ( x)   A ( x  x p )
 0 r
eN D
En ( x ) 
( x  xn )
 0 r
2

eN A  x
eN
V p ( x) 
 x p x  Vn ( x)  D

 0 r  2
 0 r

e
V  V p  Vn 
( xn2 N D  x 2p N A )
2 0 r
NA>>ND
n
E
x
V

xn2 
 xn x  
2

-xp
xn
x
Reverse biased p-n junction (II)
2 0 rV
d  xn  x p 
e
 1
1 



 ND N A 
Depletion voltage: voltage
necessary to deplete all the
junction thickness
How to know the depletion voltage of a diode?
Measurement of the capacitance
dQ dQ dX
C (V ) 

dV dX dV
e 0 r N D
C (V ) 
2V
Leakage current
The main sources of leakage current in a silicon sensor are:
1) Diffusion of charge carriers from undepleted regions of
the detector to the depleted region.
Generally well controlled, small contribution ~few nA/cm2
2) Thermal generation of electron-hole pairs in the
depleted regions.
Temperature dependent, contribution ~ mA/cm2
3) Surface currents depending on contamination, surface
defects from processing..
It may be the dominant contribution, but it can be reduced
processing guard rings
p-n junction as detector
Metal contact
photon
Charged particle
P+
-V
n-type bulk
electron
hole
Energy necessary for a m.i.p.
to produce a pair of
electron-hole in Si: 3.6 eV
Energy lost by a m.i.p. in 1
mm of silicon is ~ 300 KeV.
The typical thickness of
detectors is ~300mm.
n+-type implant
A m.i.p. produces ~25000e-≈ 4fC
+V
Fabrication
N-type silicon
SiO2
n-type wafers are oxidized at 1030oC to have the whole
surface passivated.
Using photolithographic and etching techniques, windows
are created in the oxide to enable ion implantation.
Different geometries of pads and strips can be achieved
using appropriate masks.
B
As
The next step is the doping of silicon by ion implantation.
Dopant ions are produced from a gaseous source by
ionisation using high voltage.The ions are accelerated in an
alectric field to energy in the range of 10 keV-100 keV
and then the ion beam is directed to the windows in the
oxide. P+ strips are implanted with boron, while
phosphorous or arsenic are used for the n+ contacts.
B
P+
n+
An annealing process at 600oC allows partial recovery of
the lattice from the damage caused by irradiation.
Al
The next step is the metallisation with aluminium, required
to make electrical contact to the silicon. The desired
pattern can be achieved using appropriate masks.
The last step before cutting is the passivation, which
helps to maintain low leakage currents and protects the
junction region from mechanical and ambient degradation.
Detecting charged particles
•The impinging charged particles generate electron-hole pairs
ionization
•Electron and holes drift to the electrodes under the effect of the
electric field present in the detector volume.
•The electron-hole current in the detector induces a signal at the
electrodes on the detector faces.
Metal contact
P+-type implant
Charged particle
-V
n-type bulk
E
electron
hole
n+-type implant
Reverse
bias
+V
Charged particle detection
• Energy loss mainly due to ionization
– Incident particle interacts with external electrons of Si
atoms
• All charged particles ionize
• Amount of ionization depends on:
– particle velocity
– particle charge
– medium density
Minimum Ionizing Particle
K L