Paradoxes Paradoxes of steadyof steady--state and pulse operational modestate and pulse operational mode
characteristics of silicon detectorscharacteristics of silicon detectorsirradiated by ultrairradiated by ultra--high doses of high doses of γγ--raysrays
E. Verbitskaya1, Z. Li2, E. Fretwurst3, V. Eremin1, I Ilyashenko1, J. Kierstead2, C. Wilburn4, R. Röder5
1Ioffe Physico-Technical Institute of Russian Academy of Sciences, Russia2Brookhaven National Laboratory, USA
3Institute for Experimental Physics, Hamburg University, Germany4Micron Semiconductor, UK
5CiS Institute for Microsensors, Germany
1st RD50 Workshop on radiation-hard semiconductor devicesfor very high luminosity colliders
CERN, 2-4 October 2002
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
OutlineOutline1. Motivation and earlier results2. Experimental
Steady-state characteristics3. Dependences of Vfd and Neff vs. dose in detectors from standard
and oxygen-rich Si:• experimental C-V dependencies;• space charge sign verification using TCT• modeling of Neff vs. D dependencies4. Changes of I-V characteristics with dose accumulation5. Temperature dependencies of the reverse current
Characteristics of pulse operational mode6. Distortions of current pulse response shapes in detectors from standard
and oxygen-rich Si irradiated by ultra-high dose of γ-rays7. Room temperature polarization in detectors from oxygenated Si
8. DiscussionConclusions
Motivation and earlier resultsMotivation and earlier resultsHigh doses of γ- rays play a significant role in a linear collider
0
50
100
150
200
250
300
0 100 200 300 400
Dose (Mrad)
Vfd
(V) n
orm
alis
ed to
300
µm
1.1 kOhm cm, Oxygenated1.1 kOhm cm, Standard1.2-3 kOhm cm,Oxygenated1.2-3 kOhm cm, Standard
B. Dezillie et al, IEEE Trans. NS-47(2002) 1892
Standard Si: gradual Vfd reduction till the SCSI
Si oxygen-rich (Si HTLTprocessed at BNL):
no SCSI up to 575 Mrad
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
ExperimentalExperimental
BNL detectors: n-Si Wacker; ρ = 1 kOhm⋅cm; orientation <111>; Oxygenation: BNL High Temperature Long Time process
+ Thermal Donors: Si HTLT(TD)CIS detectors: n-Si Wacker; ρ = 4 kOhm⋅cm; orientation <111> and <100>;
Oxygenation: CIS Diffusion Oxygen Float Zone (DOFZ) process;oxygenation time: 24, 48, 72 h
MS detectors: n-Si Wacker; ρ = 15 kOhm⋅cm; orientation <111>;Oxygenation: MS DOFZ process + TD
Experimental technique1. C-V and I-V measurements;2. TCT measurements with a red laser (λ = 660 nm); 3. I-T measurements
Irradiation: BNL, 60Co source
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Detectors irradiated up to ultraDetectors irradiated up to ultra--high Dhigh D
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
producer # material orientation D (Mrad)
904-96 standard 339904-100 -“- 1647906-97 oxygenated + TDs 947
BNL
906-101 -“-
<111>
9471940-18-5 standard 3391940-18-6 -“- 8352015-11-5 oxygenated + TDs 1758
MS
2015-11-6 -“-
<111>
947CA0749 standard 1647CB0249 oxygenated 1758CC1549 -“- 947
CIS
CD2040 -“-
<111>
1758CE1549 standard 1260CF0449 oxygenated 947CG1049 -“- 947
CIS
CH2149 -“-
<100>
1758
SteadySteady--state characteristics of state characteristics of γγ--irradiated detectorsirradiated detectorsCC-- V characteristicsV characteristics
SiSi standard, standard, Ultra-high dose: D = 0-1.65 Grad CIS <111> CA0749
1 10 100
10
100
C (
pF)
V (Volt)
D (Mrad): 42.7 190 338 519 835 1050 1260 1647
Standard Si:• Vfd changes
non-monotonically (similar to n/p)
• “bump” in C-V characteristicsmay appear after Vfd minimumwas reached
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
f = 10 kHz
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
SCSI in detectors from standard SCSI in detectors from standard SiSi
D = 338 Mrad BNL 904-100
650 660 670 680 690 700
-3
-2
-1
0
1
2
3
4
Sig
nal (
arb.
uni
ts)
Time (ns)
V (Volt): 20 35 45 50 55 60 70 80 90 100 120 150
SCSISCSI is observed in all detectorsfrom standard Si
Di is different
Laser illuminationon the p+ side:electron collection
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
CC--V characteristics: V characteristics: SiSi oxygenatedoxygenated
Ultra-high dose: D = 0-1.76 Grad MS # 2015-11-5
1 10 100
10
f = 10 kHz
C (
pF)
V (Volt)
D (Mrad): 42.7 96.2 190 947 1160 1370 1758
Si oxygenated:Vfd increases slowly and monotonicallywith D
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
SiSi oxygenated: positive space charge detectorsoxygenated: positive space charge detectors
Ultra-high dose D = 1.76 Grad CIS <111> CB0249
400 420 440 460
-2
0
2
4
Sig
nal (
arb.
uni
ts)
Time (ns)
V (Volt): 110 120 130 140 150 170 200 230 250 300
hole collectionhole collection(laser illumination on the n+ side)
Space charge is positivepositive!
Full depletion voltage in detectors from standardFull depletion voltage in detectors from standard SiSi
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
0 200 400 600 800 1000 1200 1400 1600 18000
100
200
300
400
500
D (Mrad)
Vfd
Vfd (
Vol
t) 904-100 1940-18-6 CA0749 CE1549
Vfd as-measured:difference in thickness~5.5%
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Standard Standard SiSi: : QQ--V dependenciesV dependencies
0 100 200 300 400 500 600 700
1
Vp-off
= 487 V
VQsat
= 595 V
Q (
arb.
uni
ts)
V (Volt)
D, Mrad: 338.7 392.2 461.2 510 620.4 719.8 835 1050 1259 1647
Vp-off – pinch-off occurs
VQsat – E(x) is linear, Q-V dependence is close to saturation
Vp-off
p+ n+
V2
V1
pinch
-off
VQsat>Vfd
Standard Standard SiSi: : comparison of Ccomparison of C--V and QV and Q--V data on V data on VVfdfd
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
0 200 400 600 800 1000 1200 1400 1600 18000
100
200
300
400
500
600
16%
V
fd (V
olt)
D (Mrad)
Vfd (C-V)
Vp-off
(Q-V) V
Qsat (Q-V)
BNL 904-100
• Vfd derived from C-V is close to Vp-off
• Vfd derived from Q-Vdependencies (VQsat)
corresponds to Vat which E is closeto linear and >Vfd(C-V)
Results are similar to earlier data in detectors irradiated by p/n
[V. Eremin et al, 6th ROSE Workshop on Rad. Hard. Silicon Det., CERN/LEB 2006-006, Oct 2000, pp. 387-390]
The difference betweenVfd (C-V) and VQsat
increases with dose
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Space charge concentration NSpace charge concentration Neffeff in detectors from standard in detectors from standard SiSi
0 200 400 600 800 1000 1200 1400 1600 1800
-8
-6
-4
-2
0
2
4
Nef
f (10
12 c
m-3)
D (Mrad)
experiment: BNL <111> 904-100 MS <111> 1940-18-6 CIS <111> CA0749 CIS <100> CE1549
fit: lines
Neff = Nd0⋅exp(-γD) -βD
# orientation Nd0 (cm-3) γ (Mrd-1) β (cm-3Mrd-1) Di (Mrd)904-100 <111> 3.86⋅1012 2.47⋅10-3 4.8⋅109 325
1940-18-6 <111> 7.23⋅1011 (7.0⋅10-5)? 8.1⋅109 88CA0749 <111> 8.91⋅1011 1.88⋅10-3 2. 7⋅109 190CE1549 <100> 9.36⋅1011 7.2⋅10-4 4.5⋅109 190
γ - donor removal rateβ - acceptor introduction rate
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
VVfdfd and Nand Neffeff in detectors from standard and oxygenated in detectors from standard and oxygenated SiSi
0 200 400 600 800 1000 1200 1400 1600 1800
-6
-4
-2
0
2
Neff
Nef
f (10
12 c
m-3)
D (Mrad)
0 200 400 600 800 1000 1200 1400 1600 18000
100
200
300
400
500
Vfd
Vfd (
Vol
t)
Si stand., 1940-18-6 Si oxyg., 2015-11-6 -"- 2015-11-5
In contrast to standard Siin oxygenated detectorsVfd increases slowlyand monotonically with dose
Positive space chargeis accumulated with Dup to ultra-high dose of 1.75 Grad
Donor type defect introductionin oxygenated Si is uniquefor γ-irradiation
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Modeling NModeling Neffeff vs. D dependencies in detectors from oxygenatedvs. D dependencies in detectors from oxygenated SiSi
0 200 400 600 800 1000 1200 1400 1600 1800 20000.0
0.5
1.0
1.5
2.0
2.5
Nef
f (10
12 c
m-3)
D (Mrad)
experiment: CIS CB0249 CIS CD2049 CIS CH2149 MS 2015-11-5
fit: CB0249 (0-1.76 Grd) -"- (0-1.37 Grd) CD2049 CH2149 2015-11-5
# oxygen.time (h)
dose rangeGrad
Nd0cm-3
Nd1cm-3/Mrd
m
2015-11-5 <111> ? 0-1.76 1.625⋅1012 1.48⋅107 1.449CB0249 <111> 24 0-1.76 1.13⋅1012 5.24⋅104 2.27CB0249 <111> 24 0-1.37 1.09⋅1012 1.52⋅107 1.47CD2049 <111> 72 0-1.76 1.038⋅1012 2.54⋅108 1.125CH2149 <100> 72 0-1.76 7.89⋅1011 7.89⋅107 1.266
Model: Nd0 + Nd1⋅Dm -
• donor type defect introduction• superlinear dependence on D
The power in superlinear dependence decreases withoxygenation time?
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
II--V V characteristicscharacteristics
1 10 100
10-3
10-2
10-1
100
101
pinch-off
SCSI
I (µA
)
V (Volt)
D (Mrad): 0 42.7 190 284 392.2 720 1260 1647
1 10 100
10-2
10-1
100
I (µA
)
V (Volt)
D (Mrad): 42.7 96.2 190 392.2 947 1160 1370 1760
SiSi oxygenatedoxygenated, MS 2015, MS 2015--1111--55Si standard, BNL 904-100
• Current values are different at the same D• Standard Si, beyond SCSI: at V<Vfd edge current on the n+-side,different from stabilization of the reverse current in detectorsas-irradiated by high F of neutrons/protons
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Reverse current density on dose dependenciesReverse current density on dose dependencies
0 200 400 600 800 1000 1200 1400 1600 18000.0
0.5
1.0
1.5
2.0
2.5
I fd (µ
A)
D (Mrad)
Si stand., CA0749 Si oxygenated
CB0249 CC1549 CD2049
Comparison of standard and oxygenatedComparison of standard and oxygenated SiSi
0 200 400 600 800 1000 1200 1400 1600 18000.0
0.5
1.0
1.5
2.0
sublinearlinear:
(1-1.5)x10 -15 A/cm2rd
jfd(D)
j fd (µ
A/c
m2 )
D (Mrad)
MS 2015-11-5 CB0249 <111> CD2049 <111> CH2149 <100> linear fit:
2015-11-5
Oxygenated Oxygenated SiSi
Standard Si:reverse current density at Vfd shows
superlinear increase with dose
Oxygenated Si:D < 1 Grad: linear increasedamage coefficient of 10-15A/cm2⋅radD > 1 Grad: sublinear increase
j at Vfd
stand.
oxygen.
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Reverse current vs. temperature dependencesReverse current vs. temperature dependences
0.0035 0.0040 0.0045 0.0050
10-4
10-3
10-2
10-1
100
101
stand.
oxygen
j (µA
/cm
2 )
T-1 (K-1)
Micron Semicond. Si stand., 1940-18-6; 835 Mrd Si oxyg., 2015-11-6; 947 Mrd
CIS <111>: Si stand., CA0749; 16470 Mrd Si oxyg., CC1549; 957 Mrd Si oxyg., CB0249; 1759 Mrd
D: 0.84-1.76 GradT: 295 to 200 KBulk generation currentis analyzed: I at V >I at V > VVfdfd
Model: Shockley-Read-Hall statistics
I-T analysis: [E. Verbitskaya et al., 5th ROSE Workshop, March 2000, CERN LEB 2000-005, p.300]
Si standard: various slope at different TEffective generation level: Ej = 0.64-0.68 eVtwo gen. levels: midgap level ~0.57 eV (lower T); and Ej ≈ 0.8 eV
Si oxygenated: single slope, Ej = 0.77-0.79 eV
Detector characteristics in pulse operational modeDetector characteristics in pulse operational mode
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
400 410 420 4300
5
10
15
20D = 519 Mrad
laser illuminationof the p+ side
Sig
nal (
arb.
uni
ts)
Time (ns)
V (Volt): 110 115 120 125 130 135 140 145 150 160 170 180 200 220
400 420 440 460
-2
0
2
4
Sig
nal (
arb.
uni
ts)
Time (ns)
V (Volt): 110 120 130 140 150 170 200 230 250 300
Si oxygenated CB0249
electron collectionelectron collection
hole collectionhole collection
D = 1.76 Grad
Si standard BNL 904-100
DJ effect –– non-uniform double peak E(x)in heavily irradiated detectors originated from free carrier trapping from detector reverse current [V. Eremin et al, NIM A476 (2002) 556-564]
• no initial fast rise of response• abnormal increase of the slope
• time dependent Neff (+Neff ↑)• E at the n+ contact drops to zero
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Room temperature polarization in detectors from oxygenated Room temperature polarization in detectors from oxygenated SiSi
• Initially: collection in fully depleted detector, E > 0• Holes are trapped during drift and increase positive Neff
• difference between Emin and Emax increases in time• pulse response degradation occurs due to polarization
640 650 660 670 680 690
0
5
10
15
in darkness
laser illuminationof n+ side
15 min
12 min
sign
al (
arb.
uni
ts)
t (ns)
D = 1.76 Grad; V = 300 V (V ≈ 2 Vfd)
hole collection: degradation of pulse response
0 200 400 600 800 1000 12000.0
0.2
0.4
0.6
0.8
1.0
Q n
orm
aliz
ed
t (s)
V (Volt): 150 200 300
0 200 400 600 800 1000 12000.0
0.2
0.4
0.6
0.8
1.0
Q n
orm
aliz
ed
t (s)
Dependence of CCE degradation on reverse bias voltageDependence of CCE degradation on reverse bias voltage
CCEh degradation is sharp• E (at the n+ contact) drops to 0,• neutral base region evolves at n+
• holes can’t be collected• the higher is V, the larger is top
hole collection
CCEe degradation is gradual• CCE ~ w/d,• initially: w = d,• at t>top gradual reduction of w
electron collection
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
at V fixed top is thesame
for e and h
640 650 660 670 680 690
0
5
10
15
0 100 200 300 400 500 6000.0
0.2
0.4
0.6
0.8
1.0
electron drift:
laser illuminationof p+ side
Q n
orm
aliz
ed
t (s)
laser switched on in all measurem
laser switched off between the measurem.
colored lines: bias on, laser onblack lines: bias on, laser off30 s between measurements
sign
al (
arb.
uni
ts)
t (ns)
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
Polarization controlled by trapping of thermally generated holesPolarization controlled by trapping of thermally generated holes
detector is permanently biased;• laser is switched on and off -
no difference in top, no influencefrom laser generated e/h
Holes are trapped from detector reverse current
Polarization: observed earlier in Sidetectors irrad. by neutrons and operated at cryogenic T[B. Dezillie et al, NIM A452 (2000) 440]
Paradox for γ-irradiated detectors:• polarization is observed at RT;
• hole trapping in SCR with +Neff
Possible reason for RT polarization:new defects in oxygenated Si induced by γ-rays:Ej = 0.8 eV, larger τdetr
Electron collection
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
DiscussionDiscussion 1.1. Paradoxes of Paradoxes of γγ--irradiatedirradiated SiSi detectorsdetectors
• Defects responsible for radiation damage and different introduction rate?• Oxy-Si: DD or SD? ∆Neff = (0.8-1.1) ⋅1012 cm-3; NSD = 3⋅1011 cm-3 (low T data)
Si standard Si oxygenated
PPaarraaddooxx: positivesuperlinear: Nd0 + Nd1⋅D
m
m = 1.1-1.5Neff
Neff = Nd0⋅exp(-γD) -βD
negative after SCSI
linear: β ~ 109 cm-3Mrd-1 D > 0.6 Grd:linear
β+ ~ 108 cm-3Mrd-1
sublinearI (Vfd):
PPaarraaddooxxsuperlinear D < 0.7 Grd:
linear:α ~ 10-9
A/cm2⋅MrdPulse
operationmode
D > 0.5 Grd:
distortions due to DJ effect
PPaarraaddooxx:- RT polarization at D ≥ 1 Grd;
- hole trapping in SCR with +Neff
DiscussionDiscussion
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
2. Radiation hardness improvement in oxygenatedRadiation hardness improvement in oxygenatedSiSi detectorsdetectors
Vfd0 – full depletion voltage before irradiation;VfdF – full depletion at the range border (D = 0.95 or 1.65 Grad)jF – reverse current density at VfdF;R = VfdF/Vfd0 – Vfd increase ratio within the dose range;GGFF == RRstst//RRoxyoxy - gain in thegain in the VVfdfd reductionreduction in oxygenated detectors with respect to detectors from standard Si irradiated to the same border dose
* Vfd0 is smaller for CIS <100>
The difference in GFand current ratio betweenthe dose ranges arises from different dependenciesof Vfd and j on D
# range,Grad
GF jF st/ jF oxy
MS 0-0.84 22.5 4.4CIS
<111>0-0.950-1.75
2.0* 2.7
3.64.6
CIS<100>
0-0.950-1.75
3* 4
3.23.3
E. Verbitskaya et al, 1st RD50 Workshop, CERN, 2-4 October 2002
ConclusionsConclusions
1. The radiation hardness improvement of detectors from oxygenated Si irradiated by γ-rays actually extends now up toultra-high dose range of 1.75 Grad
2. The gain in the Vfd reduction in oxygen rich detectors with respect to detectors from standard Si is ~2 in the range of 1 Grad and increases to ~3 in the ultra-high dose range 1-1.75 Grad
3. ParadoxParadox of oxygenated Si: • inhibited degradation of steady-state characteristics even at ultra-high dose;• collapse at D ~ 1 Grad due to RT polarization in pulse operational mode!