Comparing Si and Diamond for SLHC

Folklore up to a few years ago:p y g

Diamond: radiation hard up to 1016 n/cm2Diamond: radiation hard up to 1016 n/cm2

Si radiation hard up to 1014 n/cm2

Diamond available soon as largeDiamond available soon as large single crystal wafersallowing MIP detection

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 1

Comparing Si and Diamond for SLHCFolklore in 2010:

Si usable up to a few 1016 n/cm2(albeit large leakage current, i.e. needsstrong cooling)g g)

Diamond single crystals in large quantitiesDiamond 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, 14.12.2009 2

(but small leakage current, no cooling)

Why 10^16 needed?

12

2 s-1

]

R.Garoby - LHCC - July 2008 - "Upgrade Plans for the CERN Accelerator Comples"R.Garoby - LHCC - July 2008 - "Upgrade Plans for the CERN Accelerator Comples"FZimmermann Feb2009 "SLHCMachinePlans"FZimmermann Feb2009 "SLHCMachinePlans"

SUPER - LHC (5 years, 2500 fb-1)Pixel (?) Ministrip(?)

8

10

034 c

m-2

no PHASE II upgrade no PHASE II upgradeincluding PHASE II upgradeincluding PHASE II upgrade

F.Zimmermann- Feb. 2009 - SLHC MachinePlansF.Zimmermann- Feb. 2009 - SLHC MachinePlans

Ne5

1016Pixel (?) Ministrip(?)

Macropixel (?)

6

8

osity

[10 New

injectors + IR upgrade phase 2

Linac4 + 5

1015

cm-2

] total fluence Φeqtotal fluence Φeq

2

4

Lum

ino IR

upgrade phase 1Colli

matio 1014

5

Φeq

[

neutrons Φeq

pions Φeq

8 0 2 4 6 8 0 2 4 60

2

Peak

L

[M.Moll][M.Moll]

n phase 2 5

10 q

other charged hadrons ΦeqATLASSCT-barrelATLASPixel

2008

2010

2012

2014

2016

2018

2020

2022

2024

2026

Year0 10 20 30 40 50 60

r[cm]

1013eqATLASSCT barrelATLASPixel

(microstrip detectors)

[M.Moll, simplified, scaled from ATLAS TDR]

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 3

r [cm]From M. Moll, CERN, RD50

RD50 Collaboration on radiation hardness

247 M b f 47 I tit t

38 European institutes

247 Members from 47 Institutes

pBelarus (Minsk), Belgium (Louvain), Czech Republic (Prague (3x)), Finland (Helsinki, Lappeenranta), Germany (Dortmund, Erfurt, Freiburg, Hamburg, Karlsruhe, Munich), Italy (Bari, Fl P d P i Pi T t ) Lith i (Vil i )Florence, Padova, Perugia, Pisa, Trento), Lithuania (Vilnius), Netherlands (NIKHEF), Norway (Oslo (2x)), Poland(Warsaw(2x)), Romania (Bucharest (2x)), Russia (Moscow, St.Petersburg), Slovenia (Ljubljana), Spain (Barcelona, Valencia), g), ( j j ), p ( , ),Switzerland (CERN, PSI), Ukraine (Kiev), United Kingdom (Glasgow, Lancaster, Liverpool)

8 North-American institutesCanada (Montreal), USA (BNL, Fermilab, New Mexico,Purdue, Rochester, Santa Cruz, Syracuse)

1 Middle East instituteIsrael (Tel Aviv)

D t il d b li t htt // h/ d50

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 4

Detailed member list: http://cern.ch/rd50

Diamond radiation hardness decreases at low energies

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 5

NIEL mainly by low energy impinging particles

Above 100 MeValmost 100%ionization lossonly 10% in ionizing losses, rest mainlydi l t f l ildisplacements from nuclear recoils, i.e. Non-Ionizing Energy Losses (NIEL)

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 6

Radiation damage by nuclear recoils in Si and C

Z Ions NIEL14 417 4.213 910 9.0612 1384 12.47

Mg, Ali t t11 1021 8.86

10 1225 8.459 265 1.418 493 2.09

important recoils in Si

7 398 1.316 909 2.365 270 0.554 383 0.663 662 0.672 11152 4.41 46107 0.9Total 6559 57.38

10 GeV protonsZ Ion NIEL6 698 0.85 869 0.774 584 0.443 1133 0.552 10625 2.01

Displacement defects calculated with SRIM.COM only Heimportant recoil in C

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 7

1 30465 0.24Total 44374 4.81

Comparison with NIEL cross sections

NIEL in Silicon NIEL in Diamond

Wi d B J h B l Al d F i St ff M ll Ch i ti S dWim de Boer, Johannes Bol, Alexander Furgeri, Steffen Muller, Christian Sander(Karlsruhe U.) , Eleni Berdermann, Michal Pomorski (Darmstadt, GSI) , Mika Huhtinen (CERN) .e-Print: arXiv:0705.0171, Phys.Status Solidi 204:3009,2007

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 8

Hadron spectra at the LHC

Hadrons below 100 MeVnon-negligible, evenat LHCat LHC

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 9

Expected Signal degradation for LHC Silicon Sensors

Note: Measured partly under different conditions! Lines to guide the eye (no modeling)!( g)

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 10

Observed charge after 1016 n/cm2 much larger than expected!

• CCE: ~7300e (~30%)after ~ 1×1016cm-2 800V

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 11

Use n-in-p detectors for high fluences

Present LHC strip detectorsFuture SLHC strip detectors ??

will invert from n to pwill invert from n to pafter radiation

non-inverted, under-depleted: inverted to “p-type”, under-depleted:

•Limited loss in CCE

•Less degradation with under-depletion

p yp p

• Charge spread – degraded resolution

• Charge loss – reduced CCE

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 12

Radiation damage introduces acceptors

B B B B B After high fluences the

hh h hhAfter high fluences theacceptor concentrationbecomes so high, that thedetector cannot be depleted

hB

detector cannot be depletedanymore and results instrong negative spacecharge, i.e. high E-field. h

hh

B

h

g , g

hh h

P P PP P E

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 13

+

Drift of ionized cloud reveals electric field (TCT measurement)

1,2U= -40V

1,4U= -260VU 280V

Φ = 0 Φ = 6·1013

0,8

1 U= -120VU= -160VU= -200V 1

1,2 U= -280VU= -300VU= -320VU= -340V

0,60,6

0,8

U 340VU= -360VU= -380VU= -400V

0,2

0,4

0 2

0,4

,

085 90 95 100 105 110

0

0,2

85 90 95 100 105 110

-0,4

-0,2

-0,4

-0,285 90 95 100 105 110

Time in ns Time in ns

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 14

Electric field evolution with fluence

e

1.6E+05

1.8E+05

F 1 14

N on P silicon detector(single peak model)

V. Eremin, RD50, Nov. 2009

n po

ssib

le

1.2E+05

1.4E+05

m

F=1e14F=3e14F=1E15F=3E15

(single peak model)

iplic

atio

n

1.0E+05

c fie

ld, V

/cm F 3E15

F=1E16

d = 300umV = 500V ar

ge m

ulti

6.0E+04

8.0E+04

Elec

tric V 500V

g = 1.7e-2 cm-1 Cha

2.0E+04

4.0E+04

0.0E+000.0E+00 5.0E-03 1.0E-02 1.5E-02 2.0E-02 2.5E-02 3.0E-02

Distance cm

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 15

Distance, cm

How does charge multiplication work?

Avalanche in Silicon:

holes produced by impact ionization in high field near strip drift back in volume, p ,

recombine with electron and emitted

photon generates „Auger“ electrons, which drift to electrode and process repeats so charge multiplication is obtainedand process repeats, so charge multiplication is obtained.

A large enough hole current reduces field on hit strip, g g p,thus stopping avalanche

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 16

Applications of avalanche mechanism: Zener diodes, Si-photomultipliers, ….

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 expected3) TCT signal not saturated at high voltages, as expectedin saturated drift region is reached

4) TCT signal delayed if holes are produced at strip) g y p preceiving electrons

5) TCT signal and leakage current BOTH increases with bias voltage

6) Lorentz angle may change sign, if holes are produced sufficiently stronglysufficiently strongly

Evidence 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 them!

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 17

( j j ) g, , g pFirst hint for 5) measured in Karlsruhe (Diplomarbeit M. Schneider)

Charge collection above 100%Charge collection above 100%High CCE measured from different groups with silicon strip detectors at high fluences and high bias voltages (L’pool, Ljubljana). Device modeling using Device modeling using extrapolated parameters from low extrapolated parameters from low fluencefluence region failsregion fails completely!completely!p pp p gg p yp y

I. Mandić et al., RESMDD 08, Firenze, 2008.

full charge collection (140 μm)(140 μm)

StripsT. Affolder et al., IEEE NSS-MIC, Orlando, 2009.

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!

K b RD50 N 2009

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 181812/19/200

The work presented in this work is submitted to IEEE-TNS. Kramberger, RD50, Nov. 2009

“Edge“Edge--TCTTCT””

IR beam penetrates whole detector from side

Advantages:• Position of e-h generation can be controlled by moving tables• the amount of injected e-h pairs can be controlled by tuning the laser powerthe amount of injected e h pairs can be controlled by tuning the laser power• easier mounting and handling• not only charge but also induced current is measured – a lot more information is obtainedDrawbacks:

A li bl l f t i / i l d t t if 1060 l i d (li ht t t t d i• Applicable only for strip/pixel detectors if 1060 nm laser is used (light must penetrate guard ring region)

• Only the position perpendicular to strips can be used due to widening of the beam! Beam is “tuned” for a particular strip Li ht i j ti id h t b li h d t h d f d th l ti

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 191912/19/200

• Light injection side has to be polished to have a good focus – depth resolution• It is not possible to study charge sharing due to illumination of all strips

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 trappingestimate electron trapping times:

psttyI pp 670exp

)ns69.2,m175( 22 =→⎟⎟⎞

⎜⎜⎛ Δ

≈==

τμ

τeff,e~600 ps in good

pstyI eeff

eeffp

670exp)ns16.2,m125( ,

,2

=→⎟⎟⎠

⎜⎜⎝−≈

==τ

τμ

eff,eagreement with measurements of effective trapping times!

From short decay of I(y=25 μm) one can conclude that τeff,h is short (in

pp g

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 20

ff,700 ps holes drift 50-60 μm. At y<100 μm the field is present)

Velocity and electric field profiles do not give consistent picture if number of drifting carriers does not increase in some parts of the detector.

Arrival of electrons used for DPMPrompt current method – velocity profile

saturated velocity

•Lower fields at the strip side for low voltages•Significant field in the detector at moderate voltages – E>0.33 V/μm for 600 V. •Due to saturation of velocity the determination of the field becomes impossible, but the

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 21G. Kramberger,Investigation of electric field and evidence of charge 2112/19/200

signal still rises at small y!

The peak in the initial current at y=30 μm is prolonged at higher voltages. Drift of multiplied holes prolongs the signal.

The shift of the signal is of the same order as:vsat,e·tshift=30 μm sat,e shiftthe time needed for electrons to get to the strips!

]Ω

I[ 50

V/

-0.1

0

I[

-0.3

-0.2

-0.1

[ns]p[V],tbiasV700 , 1.11650 , 1.17For detector at 5·1014 cm-2 the shift does not

0 0.5 1 1.5 2

-0.5

-0.4

650 , 1.17600 , 1.18550 , 1.15500 , 1.17450 , 1.16400 , 1.15350 , 1.15300 , 1.14250 , 1.14200 , 1.16

For detector at 5·10 cm the shift does not exist once the drift velocity saturates!

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 2212/19/200

time [ns]0 0.5 1 1.5 2250 , 1.14200 , 1.16150 , 1.23100 , 1.51

Charge collection profile and Qmip correlation with current

Expected bulk current forExpected bulk current for fully depleted detector

Leakage current and <Q> are correlated!

High bias voltage significantly improves charge collection!p g

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 23

<Q> vs. bias Detector: Micron FZ Si p type micro strip Φ=5x1015 cm-2 T= 20°CDetector: Micron, FZ-Si p-type micro-strip, Φ=5x1015 cm 2, T=-20°C

( )dyyQQWy

∫=

=1 ( ) yyQ

WQ

y∫=0

(1st proof that this is not a

setup artifact)forward setup artifact)

reverse

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 24

Principle of Lorentz angle measurements

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 25

Lorentz shift becomes negative in irradiated sensors

ExB ExBExB

EExB

Irradiated detector:Non-irradiated detector:Lorentz shift to left

voltage reduction on stripwith charge multiplication->Lorentz shift to right possible

The hole current reduces field on hit strip, which creates horizontal electric field components from neighbouring strips.

In a magnetic field this can enhance charge multiplication on one side of stripmultiplication on one side of strip,

thus pulling charge sharing to „wrong“ side,

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 26

p g g g „ g ,i.e. effectively a negative Lorentz shift,

Observed Lorentz shifts

unirradiated 1015n/cm2

BB

BB B1015n/cm2 1016n/cm2

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 27

Future

Quantization of the charge multiplication and its dependence Q g p pon voltage, fluence and annealing

Building a new device modelBuilding 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 S/N UNDERWAY.STABILITY AND S/N UNDERWAY.

MANY INTERESTING RESULTS EXPECTED TO COME

Wim de Boer, Karlsruhe CARAT Workshop, GSI, 14.12.2009 28