Post on 27-Feb-2021
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 1
Comparing Si and Diamond for SLHC
Folklore up to a few years ago
Diamond radiation hard up to 1016 ncm2
Si radiation hard up to 1014 ncm2
Diamond available soon as large single crystal wafersallowing MIP detection
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 2
Comparing Si and Diamond for SLHCFolklore in 2010
Si usable up to a few 1016 ncm2(albeit large leakage current ie needsstrong cooling)
Diamond single crystals in large quantitiesat reasonable prices still difficultand not so radiation hard(but small leakage current no cooling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 3
Why 10^16 needed20
0820
1020
1220
1420
1620
1820
2020
2220
2420
26
Year
0
2
4
6
8
10
12
Peak
Lum
inos
ity [1
034 c
m-2
s-1]
no PHASE II upgrade no PHASE II upgradeincluding PHASE II upgradeincluding PHASE II upgrade
RGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesRGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesFZimmermann - Feb 2009 - SLHC Machine Plans FZimmermann - Feb 2009 - SLHC Machine Plans
[MMoll][MMoll]
New injectors + IR upgrade phase 2
Linac4 + IR upgrade phase 1
Collimation phase 2
0 10 20 30 40 50 60r [cm]
1013
51014
51015
51016
Φeq
[cm
-2] total fluence Φeqtotal fluence Φeq
neutrons Φeq
pions Φeq
other charged
SUPER - LHC (5 years 2500 fb-1)
hadrons ΦeqATLAS SCT - barrelATLAS Pixel
Pixel () Ministrip ()Macropixel ()
(microstrip detectors)
[MMoll simplified scaled from ATLAS TDR]
From M Moll CERN RD50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 4
8 North-American institutesCanada (Montreal) USA (BNL Fermilab New MexicoPurdue Rochester Santa Cruz Syracuse)
1 Middle East instituteIsrael (Tel Aviv)
38 European institutesBelarus (Minsk) Belgium (Louvain) Czech Republic (Prague (3x)) Finland (Helsinki Lappeenranta) Germany (Dortmund Erfurt Freiburg Hamburg Karlsruhe Munich) Italy (Bari Florence Padova Perugia Pisa Trento) Lithuania (Vilnius) Netherlands (NIKHEF) Norway (Oslo (2x)) Poland(Warsaw(2x)) Romania (Bucharest (2x)) Russia (Moscow StPetersburg) Slovenia (Ljubljana) Spain (Barcelona Valencia) Switzerland (CERN PSI) Ukraine (Kiev) United Kingdom (Glasgow Lancaster Liverpool)
RD50 Collaboration on radiation hardness
247 Members from 47 Institutes
Detailed member list httpcernchrd50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 5
Diamond radiation hardness decreases at low energies
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 2
Comparing Si and Diamond for SLHCFolklore in 2010
Si usable up to a few 1016 ncm2(albeit large leakage current ie needsstrong cooling)
Diamond single crystals in large quantitiesat reasonable prices still difficultand not so radiation hard(but small leakage current no cooling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 3
Why 10^16 needed20
0820
1020
1220
1420
1620
1820
2020
2220
2420
26
Year
0
2
4
6
8
10
12
Peak
Lum
inos
ity [1
034 c
m-2
s-1]
no PHASE II upgrade no PHASE II upgradeincluding PHASE II upgradeincluding PHASE II upgrade
RGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesRGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesFZimmermann - Feb 2009 - SLHC Machine Plans FZimmermann - Feb 2009 - SLHC Machine Plans
[MMoll][MMoll]
New injectors + IR upgrade phase 2
Linac4 + IR upgrade phase 1
Collimation phase 2
0 10 20 30 40 50 60r [cm]
1013
51014
51015
51016
Φeq
[cm
-2] total fluence Φeqtotal fluence Φeq
neutrons Φeq
pions Φeq
other charged
SUPER - LHC (5 years 2500 fb-1)
hadrons ΦeqATLAS SCT - barrelATLAS Pixel
Pixel () Ministrip ()Macropixel ()
(microstrip detectors)
[MMoll simplified scaled from ATLAS TDR]
From M Moll CERN RD50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 4
8 North-American institutesCanada (Montreal) USA (BNL Fermilab New MexicoPurdue Rochester Santa Cruz Syracuse)
1 Middle East instituteIsrael (Tel Aviv)
38 European institutesBelarus (Minsk) Belgium (Louvain) Czech Republic (Prague (3x)) Finland (Helsinki Lappeenranta) Germany (Dortmund Erfurt Freiburg Hamburg Karlsruhe Munich) Italy (Bari Florence Padova Perugia Pisa Trento) Lithuania (Vilnius) Netherlands (NIKHEF) Norway (Oslo (2x)) Poland(Warsaw(2x)) Romania (Bucharest (2x)) Russia (Moscow StPetersburg) Slovenia (Ljubljana) Spain (Barcelona Valencia) Switzerland (CERN PSI) Ukraine (Kiev) United Kingdom (Glasgow Lancaster Liverpool)
RD50 Collaboration on radiation hardness
247 Members from 47 Institutes
Detailed member list httpcernchrd50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 5
Diamond radiation hardness decreases at low energies
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 3
Why 10^16 needed20
0820
1020
1220
1420
1620
1820
2020
2220
2420
26
Year
0
2
4
6
8
10
12
Peak
Lum
inos
ity [1
034 c
m-2
s-1]
no PHASE II upgrade no PHASE II upgradeincluding PHASE II upgradeincluding PHASE II upgrade
RGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesRGaroby - LHCC - July 2008 - Upgrade Plans for the CERN Accelerator ComplesFZimmermann - Feb 2009 - SLHC Machine Plans FZimmermann - Feb 2009 - SLHC Machine Plans
[MMoll][MMoll]
New injectors + IR upgrade phase 2
Linac4 + IR upgrade phase 1
Collimation phase 2
0 10 20 30 40 50 60r [cm]
1013
51014
51015
51016
Φeq
[cm
-2] total fluence Φeqtotal fluence Φeq
neutrons Φeq
pions Φeq
other charged
SUPER - LHC (5 years 2500 fb-1)
hadrons ΦeqATLAS SCT - barrelATLAS Pixel
Pixel () Ministrip ()Macropixel ()
(microstrip detectors)
[MMoll simplified scaled from ATLAS TDR]
From M Moll CERN RD50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 4
8 North-American institutesCanada (Montreal) USA (BNL Fermilab New MexicoPurdue Rochester Santa Cruz Syracuse)
1 Middle East instituteIsrael (Tel Aviv)
38 European institutesBelarus (Minsk) Belgium (Louvain) Czech Republic (Prague (3x)) Finland (Helsinki Lappeenranta) Germany (Dortmund Erfurt Freiburg Hamburg Karlsruhe Munich) Italy (Bari Florence Padova Perugia Pisa Trento) Lithuania (Vilnius) Netherlands (NIKHEF) Norway (Oslo (2x)) Poland(Warsaw(2x)) Romania (Bucharest (2x)) Russia (Moscow StPetersburg) Slovenia (Ljubljana) Spain (Barcelona Valencia) Switzerland (CERN PSI) Ukraine (Kiev) United Kingdom (Glasgow Lancaster Liverpool)
RD50 Collaboration on radiation hardness
247 Members from 47 Institutes
Detailed member list httpcernchrd50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 5
Diamond radiation hardness decreases at low energies
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 4
8 North-American institutesCanada (Montreal) USA (BNL Fermilab New MexicoPurdue Rochester Santa Cruz Syracuse)
1 Middle East instituteIsrael (Tel Aviv)
38 European institutesBelarus (Minsk) Belgium (Louvain) Czech Republic (Prague (3x)) Finland (Helsinki Lappeenranta) Germany (Dortmund Erfurt Freiburg Hamburg Karlsruhe Munich) Italy (Bari Florence Padova Perugia Pisa Trento) Lithuania (Vilnius) Netherlands (NIKHEF) Norway (Oslo (2x)) Poland(Warsaw(2x)) Romania (Bucharest (2x)) Russia (Moscow StPetersburg) Slovenia (Ljubljana) Spain (Barcelona Valencia) Switzerland (CERN PSI) Ukraine (Kiev) United Kingdom (Glasgow Lancaster Liverpool)
RD50 Collaboration on radiation hardness
247 Members from 47 Institutes
Detailed member list httpcernchrd50
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 5
Diamond radiation hardness decreases at low energies
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 5
Diamond radiation hardness decreases at low energies
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 6
NIEL mainly by low energy impinging particles
Above 100 MeValmost 100ionization loss
Diamon
d
Silic
on
only 10 in ionizing losses rest mainlydisplacements from nuclear recoils ie Non-Ionizing Energy Losses (NIEL)
Lindh
ard pa
rtitio
n func
tion
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 7
Radiation damage by nuclear recoils in Si and C
Z Ions NIEL14 417 4213 910 90612 1384 124711 1021 88610 1225 8459 265 1418 493 2097 398 1316 909 2365 270 0554 383 0663 662 0672 11152 441 46107 09Total 6559 5738
10 GeV protonsZ Ion NIEL6 698 085 869 0774 584 0443 1133 0552 10625 2011 30465 024Total 44374 481
Mg Alimportantrecoils in Si
only Heimportantrecoil in C
Displacement defects calculated with SRIMCOM
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 8
Comparison with NIEL cross sections
NIEL in Silicon NIEL in Diamond
Wim de Boer Johannes Bol Alexander Furgeri Steffen Muller Christian Sander (Karlsruhe U) Eleni Berdermann Michal Pomorski Darmstadt GSI ( ) Mika Huhtinen (CERN) e-Print arXiv07050171 PhysStatus Solidi 20430092007
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 9
Hadron spectra at the LHC
Hadrons below 100 MeVnon-negligible evenat LHC
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 10
Expected Signal degradation for LHC Silicon Sensors
Note Measured partly under different conditions Lines to guide the eye (no modeling)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 11
Observed charge after 1016 ncm2 much larger than expected
bull CCE ~7300e (~30)after ~ 1times1016cm-2 800V
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 12
Use n-in-p detectors for high fluences
Future SLHC strip detectors Present LHC strip detectors
will invert from n to pafter radiation
non-inverted under-depleted
bullLimited loss in CCE
bullLess degradation with under-depletion
inverted to ldquop-typerdquo under-depleted
bull Charge spread ndash degraded resolution
bull Charge loss ndash reduced CCE
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 13
Radiation damage introduces acceptors
P P P
hh
h
h
P P
B B B B B
h
h h
B
h
h
E
After high fluences theacceptor concentrationbecomes so high that thedetector cannot be depletedanymore and results instrong negative spacecharge ie high E-field
+
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 14
Drift of ionized cloud reveals electric field (TCT measurement)
-04
-02
0
02
04
06
08
1
12
85 90 95 100 105 110
U= -40VU= -120VU= -160VU= -200V
-04
-02
0
02
04
06
08
1
12
14
85 90 95 100 105 110
U= -260VU= -280VU= -300VU= -320VU= -340VU= -360VU= -380VU= -400V
Time in ns Time in ns
Φ = 0 Φ = 6middot1013
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 15
Electric field evolution with fluence
00E+00
20E+04
40E+04
60E+04
80E+04
10E+05
12E+05
14E+05
16E+05
18E+05
00E+00 50E-03 10E-02 15E-02 20E-02 25E-02 30E-02
Distance cm
Elec
tric
fiel
d V
cm
F=1e14F=3e14F=1E15F=3E15F=1E16
d = 300umV = 500Vg = 17e-2 cm-1
N on P silicon detector(single peak model)
V Eremin RD50 Nov 2009
Cha
rge
mul
tiplic
atio
npo
ssib
le
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 16
How does charge multiplication work
Avalanche in Silicon
holes produced by impact ionization in high fieldnear strip drift back in volume
recombine with electron and emitted
photon generates bdquoAugerldquo electrons which drift to electrodeand process repeats so charge multiplication is obtained
A large enough hole current reduces field on hit strip thus stopping avalanche
Applications of avalanche mechanism Zener diodes Si-photomultipliers hellip
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 17
Evidence for charge multiplicationEvidence for charge multiplication1) Collected charge larger than in unirradiated sensor
2)Second peak in TCT after electrons reach strip
3) TCT signal not saturated at high voltages as expectedin saturated drift region is reached
4) TCT signal delayed if holes are produced at stripreceiving electrons
5) TCT signal and leakage current BOTH increases with bias voltage6) Lorentz angle may change sign if holes are produced sufficiently stronglyEvidence for 1) by Affolder for 2-5 shown by G Kramberger and M Milovanovic (Ljubljana) at RD50 meeting CERN Nov 2009 Following plots taken from themFirst hint for 5) measured in Karlsruhe (Diplomarbeit M Schneider)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 18
Charge collection above 100Charge collection above 100High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (Lrsquopool Ljubljana) Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completelycompletely
We need a new tool to identify electric field and multiplication effects Ideal tool for investigation of electric field and charge multiplication in silicon detectors (paticularly heavily irradiated) is Edge-TCT
The work presented in this work is submitted to IEEE-TNS
I Mandić et al RESMDD 08 Firenze 2008
full charge collection (140 microm)
StripsT Affolder et al IEEE NSS-MIC Orlando 2009
Kramberger RD50 Nov 2009
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 19
ldquoldquoEdgeEdge--TCTTCTrdquordquo
IR beam penetrates whole detector from side
Advantagesbull Position of e-h generation can be controlled by moving tablesbull the amount of injected e-h pairs can be controlled by tuning the laser powerbull easier mounting and handlingbull not only charge but also induced current is measured ndash a lot more information is obtainedDrawbacksbull Applicable only for strippixel detectors if 1060 nm laser is used (light must penetrate guard ring
region)bull Only the position perpendicular to strips can be used due to widening of the beam Beam is ldquotunedrdquo
for a particular strip bull Light injection side has to be polished to have a good focus ndash depth resolutionbull It is not possible to study charge sharing due to illumination of all strips
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 20
A second peak emerges in the induced current signals which is related to electron drift (it shifts when moving away from the strip)
It can only be explained by electrons entering very high field at the strips where they multiply The second peak is a consequence of holes drifting away from the strips
The change of 2nd peak amplitude can be used to estimate electron trapping times
pst
tyItyI
eeffeeff
p
p
p 670exp)ns162m125()ns692m175(
2
2
2 =rarr⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆minusasymp
==
==τ
τmicromicro
τeffe~600 ps in good agreement with measurements of effective trapping times
From short decay of I(y=25 microm) one can conclude that τeffh is short (in 700 ps holes drift 50-60 microm At ylt100 microm the field is present)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 21
Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector
Prompt current method ndash velocity profile Arrival of electrons used for DPM
saturated velocity
bullLower fields at the strip side for low voltagesbullSignificant field in the detector at moderate voltages ndash Egt033 Vmicrom for 600 V bullDue to saturation of velocity the determination of the field becomes impossible but the signal still rises at small y
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 22
The peak in the initial current at y=30 microm is prolonged at higher voltages Drift of multiplied holes prolongs the signal
time [ns]0 05 1 15 2
]Ω
I[ 50
V
-05
-04
-03
-02
-01
0
[ns]p[V]tbiasV700 111650 117600 118550 115500 117450 116400 115350 115300 114250 114200 116150 123100 151
The shift of the signal is of the same order asvsatemiddottshift=30 microm the time needed for electrons to get to the strips
For detector at 5middot1014 cm-2 the shift does not exist once the drift velocity saturates
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 23
Charge collection profile and Qmip correlation with current
Expected bulk current for fully depleted detector
High bias voltage significantly improves charge collection
Leakage current and ltQgt are correlated
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 24
ltQgt vs bias Detector Micron FZ-Si p-type micro-strip Φ=5x1015 cm-2 T=-20degC
( )dyyQW
QWy
yint=
=
=0
1
forward
reverse
(1st proof that this is not a
setup artifact)
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 25
Principle of Lorentz angle measurements
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 26
Lorentz shift becomes negative in irradiated sensors
ExBE
ExBExB
Non-irradiated detectorLorentz shift to left
Irradiated detectorvoltage reduction on stripwith charge multiplication-gtLorentz shift to right possible
The hole current reduces field on hit strip which createshorizontal electric field components from neighbouring strips
In a magnetic field this can enhance chargemultiplication on one side of strip
thus pulling charge sharing to bdquowrongldquo side ie effectively a negative Lorentz shift
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 27
Observed Lorentz shifts
unirradiated
B
B
B
B
1015ncm2 1016ncm2
1015ncm2
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME
Wim de Boer Karlsruhe CARAT Workshop GSI 14122009 28
Future
Quantization of the charge multiplication and its dependence on voltage fluence and annealing
Building a new device model
What is optimum design
Can one really operate Si detectors in charge multiplication mode
PROGRAMS TO INVESTIGATE OPTIMUM DESIGN STABILITY AND SN UNDERWAY
MANY INTERESTING RESULTS EXPECTED TO COME