CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
TCAD simulations of isolation structures TCAD simulations of isolation structures for nfor n++-on-p silicon microstrip detectors-on-p silicon microstrip detectors
Claudio PiemonteClaudio [email protected]@itc.it
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
OutlineOutline
• Why is this work needed?
• Device Simulations
- p-spray
- p-spray experience @ ITC-irst
- p-stop
- combined p-spray/p-stop
• Conclusion
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Why are simulations needed?Why are simulations needed?
3 techniques available (from n+-on-n technology):
n+ n+oxide
strip 1 strip 2
p+
p- substrate
backplane
+ + + + + + + +
electron layer
isolation structureneeded to interrupt theinversion layer betweenthe strip
S1 S2S1 S2 S1 S2
p-spray p-stop p-spray/p-stop
Each technique affects differently VBR and Cint
simulations are needed to evaluate impact
high-field regions high-field regions
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Simulated structureSimulated structure
n+ n+
strip 1 strip 2
p+
p- substrate
backplane
oxide
½ strip½ strip
2D simulations of a cross section orthogonal to the strips
Reference structure: pitch = 80m n+ width = 30m Na = 7e11cm-3
W = 300m Qox=4e11cm-2
0E+00
1E-13
2E-13
3E-13
4E-13
5E-13
6E-13
0 20 40 60 80 100Bias voltage (V)
Inte
rstri
p C
apac
itanc
e (F
/cm
)
w = 40m
30m
20m
w=40mm -> w/p=0.5w=30mm -> w/p=0.375w=20mm -> w/p=0.25
w=40m -> w/p=0.5w=30m -> w/p=0.375w=20m -> w/p=0.25
Pure geometrical Cint
(i.e. Qox=0, no isol. struct)
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
p-sprayp-spray
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Transition voltage depends on the substrate concentration
Slope depends on w/p
Np1=Np2=Np3
Breakdown (1)Breakdown (1)strip 1 strip 2
oxiden+ n+
strip 1 strip 2oxide
n+ n+
VJBR = potential difference between p-spray and strip which causes breakdownVBR = bias voltage for which we reach VJBR
High Np => High VJBR
=> high VBR
Three p-spray peak concentrations:- Np1=4e16cm-3
- Np2=8e16cm-3
- Np3=12e16cm-3
0
20
40
60
80
100
120
0 200 400 600 800Bias Voltage (V)
p-s
pra
y p
ote
ntia
l (V
)VJBR
VBR
Np = 12e16cm-3
Np = 8e16cm-3
Np = 4e16cm-3
no impact ionizationmodel enabled
0
20
40
60
80
100
120
0 200 400 600 800Bias Voltage (V)
p-s
pra
y p
ote
ntia
l (V
)VJBR
VBR
Np = 12e16cm-3
Np = 8e16cm-3
Np = 4e16cm-3
VJBR
VBR
Impact ionizationenabled
0
20
40
60
80
100
120
0 200 400 600 800Bias Voltage (V)
p-s
pra
y p
ote
ntia
l (V
)VJBR
VBR
Np = 12e16cm-3
Np = 8e16cm-3
Np = 4e16cm-3
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Breakdown (2)Breakdown (2)
VBR depends on:1) slope of Vp-spray vs VBIAS which depends on w/p and Na 2) VJBR level which depends on NP, QOX and field-plate
0
20
40
60
80
100
120
0 200 400 600 800
Bias voltage (V)
p-s
pra
y p
ote
ntia
l (V
)
1. 0e11cm-2
2. 4e11cm-2
3. 8e11cm-2
4. 12e11cm-2
5. 20e11cm-2
1
2
3
4
5
0
20
40
60
80
100
120
0 200 400 600 800
Bias voltage (V)
p-s
pra
y p
ote
ntia
l (V
)
1
2
3
4
5
1. 0e11cm-2
2. 4e11cm-2
3. 8e11cm-2
4. 12e11cm-2
5. 20e11cm-2
Np2=8e16cm-3 Np3=12e16cm-3
VBR increasesfor increasing Qox
because VJBR level increases
no isol.
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Interstrip capacitance (1) Interstrip capacitance (1)
0.0E+00
2.0E-13
4.0E-13
6.0E-13
8.0E-13
1.0E-12
1.2E-12
1.4E-12
1.6E-12
1.8E-12
2.0E-12
0 50 100 150 200
Bias voltage (V)
Inte
rstr
ip C
ap
aci
tan
ce (
F/c
m)
1. Qox= 0e11cm-2
2. 4e11cm-2
3. 8e11cm-2
1
3
Np = 4e16cm-3
geometrical value
0.0E+00
2.0E-13
4.0E-13
6.0E-13
8.0E-13
1.0E-12
1.2E-12
1.4E-12
1.6E-12
1.8E-12
2.0E-12
0 50 100 150 200
Bias (V)
Inte
rstr
ip C
ap
aci
tan
ce (
F/c
m)
1. Qox = 0e11cm-2
2. 4e11cm-2
3. 12e11cm-2
4. 20e11cm-2
5. 25e11cm-2
6. 30e11cm-2
1
6
Np = 12e16cm-3
geometrical value
Np1=4e16cm-3
Np3=12e16cm-3
Cint increases with Np
Cint decreases with Qox
p-spray completely compensated => no isolation
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
0.0E+00
5.0E-13
1.0E-12
1.5E-12
2.0E-12
2.5E-12
3.0E-12
0 100 200 300 400 500Bias voltage (V)
Inte
rstr
ip c
ap
aci
tan
ce (
F/c
m)
High-dose p-spray case
no field-plate
5m field plate - Tox=500nm
5m field plate - Tox=900nm
Interstrip capacitance (2) Interstrip capacitance (2)
n+ n+
increases Cint by:
C=0.5 *ox *(FPW*L) *Tox-1
Effect of field-plate
Additional component decreases because p-spray is progressively depleted by field-plate
Tox
FPW
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
p-spray @ irstp-spray @ irst
First n+-on-p detectors with p-spray fabricated in 2004
n+
p-spray
p-spray profile along cut-lineNet-doping conc.
dose=3e12cm-2
dose=5e12cm-2
2 p-spray doses were used(labeled as low-doseand medium-dose)
Both doses isolate before irradiation
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
p-spray @ irst - breakdownp-spray @ irst - breakdown
SMART2 - FZ - Bias Line current @ 100V
1E+1
1E+2
1E+3
1E+4
1E+5
S1 S2 S3 S4 S5 S6 S7 S8 S9 S10Sensor #
I_B
L [n
A/c
m2]
low-dose
p-spray
med-dose p-spray
3 51 2 4 8 106 7 9
Vbias=100V
- Low dose p-spray no BD- Medium dose p-spray: 50m pitch no BD 100m pitch BD
- Exception: med. dose 100m pitch no BD because of larger p+
det# pitch w/pmetal
overhang[um]
1 50 0.3 42 50 0.4 43 50 0.5 44 50 0.3 25 50 0.3 6
6 100 0.15 47 100 0.25 48 100 0.35 49 100 0.25 6
10 100 0.25 8
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
Low-dose p-spray
High-dose p-spray
- At “low” voltages cap. dominated by field-plate: - larger field plate => higher capacitance - higher dose p-spray
=> longer decay
p-spray @ irst – interstrip capacitancep-spray @ irst – interstrip capacitance
det# pitch w/pmetal
overhang[um]
1 50 0.3 42 50 0.4 43 50 0.5 44 50 0.3 25 50 0.3 6
6 100 0.15 47 100 0.25 48 100 0.35 49 100 0.25 6
10 100 0.25 8
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
p-stopp-stop
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
BreakdownBreakdown
0
10
20
30
40
50
60
70
0 100 200 300 400 500 600 700 800bias voltage (V)
p-s
top
po
ten
tial (
V)
5um 10um15um 20um30um 40um
Qox = 1e12cm-2
Qox = 2e12cm-2
Similar approach to p-spray:1. determine Vp-spray = f(VBIAS)2. determine VJBR level
S1 S2
high-field regions
increasing p-stop width
p-stop potential is higher for wide implants better narrow p-stop from the breakdown viewpoint
VJBR
VJBR
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
0.0E+00
2.0E-13
4.0E-13
6.0E-13
8.0E-13
1.0E-12
1.2E-12
1.4E-12
1.6E-12
1.8E-12
2.0E-12
0 100 200 300 400 500Bias voltage (V)
Inte
rstr
ip C
ap
aci
tan
ce (
F/c
m)
Qox=4e11cm-2
growing p-stop width
40m30m
20m15m
10m
5m
Interstrip capacitanceInterstrip capacitance
S1 S2
inversion layer acts as an extensionof the n+ strip
narrow p-stop implies higher capacitance(opposite trend with respect to breakdown)
As Qox increases, interstrip capacitance increases
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
combinedcombinedp-spray/p-stopp-spray/p-stop
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
ConceptConcept
p-spray: VBR - low before irradiation - improves for increasing QOX
Cint - improves with irradiationp-stop: VBR - high before irradiation - decreases for increasing QOX
Cint - high and deteriorates for inc. QOX
improves with QOX
deteriorates with QOX
interesting solution is to combine the previous two using:- medium dose p-spray (to have sufficiently high initial VBR)
- 20/30m wide p-stop (to have low capacitance for high QOX)
Technology: - moderated p-spray (isolation structure realized with single implantation step)
- combined p-spray/p-stop (two implantation steps)
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
BreakdownBreakdown
0
30
60
90
120
150
0 200 400 600 800 1000Bias voltage (V)
p-s
top
po
ten
tial (
V)
1
2
3
4
5
VBR1 VBR2 VBR3
VBR4VBR5
1. Qox=0
2. Qox=4e11cm-2
3. Qox=10e11cm-2
4. Qox=20e11cm-2
5. Qox=30e11cm-2
For increasing QOX:S1 S2p-stopS1 S2p-stopS1 S2p-stop
1 3 4high-field regions high-field regions high-field regions2 5
VBR first increases(typical of p-spray)than decreases (typical of p-stop)as QOX grows
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
0.0E+00
2.0E-13
4.0E-13
6.0E-13
8.0E-13
1.0E-12
1.2E-12
1.4E-12
1.6E-12
1.8E-12
2.0E-12
0 100 200 300 400 500Bias voltage (V)
Inte
rstr
ip c
ap
aci
tan
ce (
F/c
m)
1. Qox=4e11cm-2
2. Qox=8e11cm-2
3. Qox=12e11cm-2
4. Qox=20e11cm-2
1
2
3
4
Interstip capacitanceInterstip capacitance
S1 S2p-stopS1 S2p-stopS1 S2p-stop
1 3 42
Cint first decreases(typical of p-spray)than increases (typical of p-stop)as QOX grows
CERN, 14-16 November 2005RD50 workshopClaudio Piemonte
ConclusionConclusion
• simulations well reproduce both breakdown and interstrip capacitance behavior in the p-spray case
• Among the simulated isolation structures, the combined p-spray/p-stop technology seems to be the most effective (as for n+-on-n detectors)