H. F.-W. Sadrozinski, PTP, Trento, March 2011 1

Updates on Punch-through Protection

H. F.–W. Sadrozinskiwith

C. Betancourt, A. Bielecki, Z. Butko, A. Deran, V. Fadeyev, S. Lindgren, C. Parker, N. Ptak, J. Wright

SCIPP, Univ. of California Santa Cruz, CA 95064 USA

• Punch-through Protection against Large Voltages on Implants• Testing of Field-breakdown with an IR Laser• Comparison with DC measurements• F. o. M. : Bulk Resistance / Saturation Resistance• Extraction of PT parameters: DC 4-R Model • Radiation and Temperature Effects• P-Dose (spray and stop)• Role of Implant Resistance: Mitigation ?• (R-C Filter Circuit, looks ok, but work is not completed)

H. F.-W. Sadrozinski, PTP, Trento, March 2011 2

• In beam accidents, the field inside the sensor breaks down when the deposited charge makes the sensor conductive. At that point, the bias voltage can reach into the sensor bulk and can impart large voltages to the implants, while the Al readout trace of AC coupled sensors are held at ground by the readout ASIC

• This can lead to the large voltages across the coupling capacitors and breaking.

•To check this, we used IR lasers to mimic the beam loss and we indeed observed large voltages on the implants T. Dubbs et al., IEEE Trans. Nuclear Science 47, 2000:1902 – 1906.

Damage from Beam Losses to AC-coupled SSD

• The reach-through (punch-through) effect was considered an elegant and effective way to limit the voltages on the implants. Special punch-through protection (PTP) structure can be designed where the geometrical layout determines the voltage limits on the implants.J. Ellison et al., IEEE Trans. Nuclear Science, 36, 1989: 267 - 271.

• PTP structures were implemented in the p-on-n SCT sensors. • But the PTP structure in SCT sensors were shown not to guarantee protection against large voltages across the coupling capacitors when IR laser pulses were used.K. Hara, et al., Nucl. Instr. and Meth. A 541 (2005) p. 15-20.

H. F.-W. Sadrozinski, PTP, Trento, March 2011 3

PTP in Upgrade Sensors ATLAS07

No PTP structure

BZ4A BZ4B BZ4C BZ4DPTP structures

N-on-P full-size sensors and “mini’s” to investigate PTP (in Zone 4A-4D)

Y. Nobu, et al., NIMA A (2010)doi:10.1016/j.nima.2010.04.080

H. F.-W. Sadrozinski, PTP, Trento, March 2011 4

• Bias ring is held to ground • Voltage on a DC pad and/or AC pad are read via an high impedance voltage divider into pico-probe or digital scope.• DC pad reflects biasing of strips, • AC pad reflects instantaneous collected charge ( ~ depleted region)

Testing Large Implant Voltages with Laser

• Alessi IR cutting laser deposits large amounts of charge inside the detector which collapses the field(>1011 e/h pairs ~ 107 MIPs ~ 1 Rad / pulse).• Intensity given by number of laser triggers ~ 4 sec apart (we used up to 3).

Voltages on strips are large, much larger than DC VPT , comparable to bias voltage.

W51-BZ4C Laser Profile Strip 10, 200Vbias -180

-160

-140

-120

-100

-80

-60

-40

-20

0

-10 0 10 20 30 40

Distance Laser - Strip (# of Strips)V

olt

age

On

Im

pla

nt

(V)

1 L Puls

3 L Pulses

T. Dubbs et al., IEEE Trans. Nuclear Science 47, 2000:1902 – 1906.

• Laser spot ~ 10 m, but large DC voltages extend over few mm.

• Peak voltage independent of laser intensity, and AC grnd/floating

H. F.-W. Sadrozinski, PTP, Trento, March 2011 5

Vnear vs. Vbias, LP=505, LN, AC Ground

0

50

100

150

200

250

0 50 100 150 200 250 300 350

Vbias [ V ]

Vn

ear

[V

]0

50

100

150

200

250

W51-BZ4A-P4

W51-BZ4B-P10

W51-BZ4C-P16

W51-BZ4D-P16

W51-BZ2-P8

W51-BZ3-P3

Laser implant voltages near RPT vs. Bias

At high bias voltages, implant voltages for PTP structures saturate. But the ones without PTP structures do not saturate (even though DC VPT were very similar)

Can this be reconciled with the DC voltage dependence of RPT ?

DC VPT

Laser near

DC Voltagefor RPT

Laser nearLaser near

DC Voltagefor RPT

DC Voltagefor RPT

Rnear = RPT in parallel to Rbias

H. F.-W. Sadrozinski, PTP, Trento, March 2011 6

PTP Structure Effectiveness

• The effectiveness of PTP structures is determined in DC i-V measurements between the strip and the bias ring. One measures the “integral” effective resistance Reff, (Rnear)which is the bias resistor Rbias in parallel to the PTP resistor RPTP.

• The measure of the effectiveness of the PTP structure is the punch-through voltage VPT , defined as the voltage at which RPTP = Rbias, i.e. Reff, = 0.5* Rbias .

• The PT voltage VPT was measured to be a few 10’s of Volt.

W51 p-stop=4e12, Vbias=200V

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

-50 -40 -30 -20 -10 0Vtest (Volts)

Re

ff (

Oh

ms

)

W51-BZ2-P8

W51-BZ3-P3

W51-BZ4A-P4

W51-BZ4B-P10

But: VPT shows very little variation with the dimension of the PTP structure (channel length L):

L [m] VPT

BZ4A 21 20

BZ4B 21 24

BZ2 / BZ3 75 / 67 21

S. Lindgren, et al., NIM A (2010) doi:10.1016/j.nima.2010.04.094

H. F.-W. Sadrozinski, PTP, Trento, March 2011 7

Issues with DC measurements

• In DC measurements interstrip punch-through can be significant: compare Rtotal (Vtest/Itest) and Rnear @ 100V test voltage.

• In DC measurements, the neighbors are held to ground through the bias resistance, i.e. the test voltage is essentially applied between neighboring strips.

• In laser measurements much smaller voltage differences between neighbors are observed, since many strips are involved: no inter-strip PT!

Resistance Comparison @ 100V Test Voltage

0.00E+00

2.00E+04

4.00E+04

6.00E+04

8.00E+04

1.00E+05

1.20E+05

1.40E+05

1.60E+05

1.80E+05

Re

sis

tan

ce

(O

hm

)

W71-BZ2 Rtotal

W71-BZ2 Reff1

W71-BZ3 Rtotal

W71-BZ3 R1eff

W71-BZ4D Rtotal

W71-BZ4D R1eff

W71-BZ6 Rtotal

W71-BZ6 R1eff

W50-BZ2 Rtotal

W50-BZ2 R1eff

W50-BZ3 Rtotal

W50-BZ3 R1eff

W50-BZ6 Rtotal

W50-BZ6 R1eff

H. F.-W. Sadrozinski, PTP, Trento, March 2011 8

Understanding PT Structures: I-V

Drift-Diffusion

SCL, no dependence on VPT

H. F.-W. Sadrozinski, PTP, Trento, March 2011 9

PT Resistances and Voltages in SCL

( )( )

Vn ~ Vbias when is small, i.e. Rsat >> Rbulk: Not Good for PTPVn saturates to Vfb when is large, i.e. Rsat << Rbulk : Good for PTP

Investigate the dependence of PT parameters on geometry, doping levels, temperature, radiation

H. F.-W. Sadrozinski, PTP, Trento, March 2011 10

Vnear vs. Vbias, LP=505, LN+LF, AC Ground

0

50

100

150

200

250

0 50 100 150 200 250 300 350

Vbias (Volts)

Vn

ear

(Vo

lts)

0

50

100

150

200

250

W51-BZ4A-P4W51-BZ4B-P10W51-BZ4C-P16W51-BZ4D-P16W51-BZ2-P8W51-BZ3-P3W51-BZ4A-P4W51-BZ4B-P10W51-BZ4C-P16W51-BZ4D-P16W51-BZ2-P8W51-BZ3-P3

A DC “4R” Model works nicely

• After breakdown of field inside the sensor, deal with DC resistor chain onlyRPTnear = Reff (RPTnear ,Rbias )

RPTfar = RPT on the far end of the strip, RPTfar > RPTnear

Rimp = Resistance of implant 15k/cm Rbulk = Bulk Resistance

PT structures are working to clamp the voltage.Non-PT structures do not.

Important for 4R Model:Fire laser both near and far. Measure both Vnear and Vfar

Calculate all R’s and I’s and V’s

Voltages near RTP structures clamped independent of laser position

H. F.-W. Sadrozinski, PTP, Trento, March 2011 11

Bulk has Resistance of 10’s k

Rbulk vs Vbias, LF

1.0E+03

1.0E+04

1.0E+05

0 50 100 150 200 250 300

Vbias [V]

Rb

ulk

(O

hm

s)

W99-BZ1-P19 Un-irr T=22 W99-BZ1-P19 Un-irr T=-10

W51-BZ2-P8 Un-irr T=22 W18-BZ2-P17 irr T=-15

W51-BZ4D-P16 Un-irr T=22 W50-BZ4D-P22 irr T=22

W50-BZ4D-P22 irr T=0 W50-BZ4D-P22 irr T=-5

W50-BZ4D-P22 irr T=-10 W51-BZ4C-P16 Un-irr T=22

W18-BZ4C-P16 Irr T=-10 W18-BZ4C-P16 Irr T=-15

W51-BZ4B-P10 Un-irr T=22 W51-BZ4A-P4 Un-irr T=22

W50-BZ4A-P4 Irr T=22 W50-BZ4A-P4 Irr T=0

W50-BZ4A-P4 Irr T=-5 W50-BZ4A-P4 Irr T=-10

Independent of TDependent on Radiation (Vfb!)

RBulk ~> 104After depletion at 150V

Ibulk vs Vbias, LF

0.0E+00

1.0E-03

2.0E-03

3.0E-03

4.0E-03

5.0E-03

6.0E-03

7.0E-03

0 50 100 150 200 250 300

Vbias

Ibu

lk (

A)

W99-BZ1-P19 Un-irr T=22W99-BZ1-P19 Un-irr T=-10W51-BZ2-P8 Un-irr T=22W18-BZ2-P17 irr T=-15W51-BZ4D-P16 Un-irr T=22W50-BZ4D-P22 Irr T=22W50-BZ4D-P22 irr T=0W50-BZ4D-P22 irr T=-5W50-BZ4D-P22 irr T=-10W51-BZ4C-P16 Un-irr T=22W18-BZ4C-P16 Irr T=-10W18-BZ4C-P16 Irr T=-15W51-BZ4B-P10 Un-irr T=22W51-BZ4A-P4 Un-irr T=22W50-BZ4A-P4 Irr T=22W50-BZ4A-P4 Irr T=0W50-BZ4A-P4 Irr T=-5W50-BZ4A-P4 Irr T=-10

Filled symbols: un-irradiated, T=22Open symbols: irradiated, T = -15 to +22Depletion ~150 V

H. F.-W. Sadrozinski, PTP, Trento, March 2011 12

PT Structures have small Rsat (~100)

Irradiation with 70 MeV protons to 1013 p/cm2

Vnear vs Vbias, LF

0

10

20

30

40

50

60

70

80

90

0 50 100 150 200 250 300

Vbias

Vn

ea

r (V

olt

s)

W51-BZ4A-P4 Un-irr T=22

W50-BZ4A-P4 Irr T=22

W50-BZ4A-P4 Irr T=0

W50-BZ4A-P4 Irr T=-5

W50-BZ4A-P4 Irr T=-10

Saturation of voltage persists post-rad, with small increase in Vnear (Vfb !) No temperature dependence either pre- or post-rad. Need higher fluences!

Un-irr, T=22

Irr, T=-10 - 22

H. F.-W. Sadrozinski, PTP, Trento, March 2011 13

PT Structures have small Rsat (~100)

Rpt near vs Ipt near, LF

1.0E+04

1.0E+05

1.0E+06

1.0E-04 1.0E-03 1.0E-02

Ipt near (Amps)

Rp

t n

ea

r (O

hm

s)

W51-BZ4A-P4 Un-irr T=22

W50-BZ4A-P4 Irr T=22

W50-BZ4A-P4 Irr T=0

W50-BZ4A-P4 Irr T=-5

W50-BZ4A-P4 Irr T=-10

PT resistance RPT ~ 1 / IPT no T dependence no radiation dependence (Vfb)

H. F.-W. Sadrozinski, PTP, Trento, March 2011 14

Non-PT Structures have large Rsat (~30k)

Vnear vs Vbias, LF

0

20

40

60

80

100

120

140

160

180

0 50 100 150 200 250 300

Vbias

Vn

ea

r (V

olt

s)

W51-BZ2-P8 Un-irr T=22

W18-BZ2-P17 irr T=-15

Rpt near vs Ipt near, LF

1.0E+04

1.0E+05

1.0E+06

1.0E-04 1.0E-03 1.0E-02

Ipt near (Amps)

Rp

t n

ea

r (O

hm

s)

W51-BZ2-P8 Un-irr T=22

W18-BZ2-P17 irr T=-15

No Saturation of Voltage

PT resistance starts to saturate

H. F.-W. Sadrozinski, PTP, Trento, March 2011 15

Vnear vs Vbias (un-irr, T=22)

0

20

40

60

80

100

120

140

0 50 100 150 200 250 300

Vbias (V)

V n

ea

r (V

)

W99-BZ1-P19 p-spray 2e12

W51-BZ4D-P16 p-stop, 4e12

W71-BZ4D-P22 p-stop, 1e13

P-spray vs. P-stop ( 3 p-doses)

Increased p-dose increases Saturation Voltage – Flat Band Voltage Vfb

P-spray 2e12

P-stop 4e12

P-stop 1e13

H. F.-W. Sadrozinski, PTP, Trento, March 2011 16

Rpt near,far vs Vnear,far (un-irr, T=22)

1.0E+04

1.0E+05

1.0E+06

0 50 100 150 200

Vnear, far (Volts)

Rp

t n

ea

r, f

ar

(Oh

ms

)

W51-BZ3-P3 near

W51-BZ3-P3 far

W51-BZ2-P8 near

W51-BZ2-P8 near

Gate Effect in RPT(near)

Difference of “identical” near and far resistances in BZ2 and BZ3:Strong gate effect of the Poly bias resistor (c.f. FOXFET)Also: in BZ4A the Poly resistor has the best channel coverage

H. F.-W. Sadrozinski, PTP, Trento, March 2011 17

Effect of Finite Implant Resistance Rimp

Limitation of application of PTP structures: finite Rimp

Implant Voltages do not saturate at high bias voltages, if finite implant resistance Rimp isolates PTP structure from breakdown region.

Proposal with CNM Barcelona to reduce strip resistance.

Vnear vs. Vbias, LP=505, LF, AC Ground

0

50

100

150

200

250

0 50 100 150 200 250 300 350

Vbias (Volts)

Vnea

r (Vo

lts)

0

50

100

150

200

250

W51-BZ4B-P10

W51-BZ4C-P16

W51-BZ4D-P16

W51-BZ2-P8

W51-BZ3-P3

W51-BZ4A-P4

Vfar vs. Vbias, LP=505, LF, AC Ground

0

50

100

150

200

250

0 50 100 150 200 250 300 350Vbias (Volts)

Vfar

(Vol

ts)

0

50

100

150

200

250

W51-BZ4B-P10

W51-BZ4C-P16

W51-BZ4D-P16

W51-BZ2-P8

W51-BZ3-P3

W51-BZ4A-P4

Laser farLaser far

Fire laser at the far end of the 1 cm strip, measure both Vnear and Vfar

Vfar > Vnear Saturation in Vnear

No saturation in Vfar

RTP structure not effectivefor Vfar

Near end, plateaufor some PT structures

Opposite end, no plateau.

H. F.-W. Sadrozinski, PTP, Trento, March 2011 18

Reducing Implant Resistance (CNM – UCSC)• Proposed solution:

– A low effective resistance of the implant strips on a silicon sensor. A desired target value is 1.5 kOhm/cm

• Not possible to increase implant doping to significantly lower the resistance

– Instead deposition of Aluminum on top of the implant

– A layer of high-quality oxide and nitride with metal strips on top to implement the AC-coupled sensor readout.

18

H. F.-W. Sadrozinski, PTP, Trento, March 2011 19

Conclusions

• An IR cutting laser mimics the beam accident conditions: large Q, collapsing E-field.• Field breakdown size ~ 0.5 mm, with a much larger area with partial breakdown.

•The observed voltages and and voltages be explained by 4 resistor DC model : RPTP(near), RPTP(far), Rbulk , Rimplant. • Results are consistent with drift-diffusion in the SCL region • Figure of Merit for punch-through voltage saturation : Rbulk / Rsat, if large then, PT.• Saturation resistance ~ L2, 100 for PTP structures, 10 k for non-PTP structures• For PTP structures, the voltage near the laser VPTP(laser, near) saturates.

Then RPTP ~20k; I ~ 5 -10 mA. • Saturation voltage determined by flat band voltage: Radiation sensitivity, no temperature dependence

• Rimplant is very important: the present value, ~15 k/cm, can effectively isolate the collapsed field from the PTP structure, increasing voltages on the implants by 100’s of volts => need low Rimplant! Work with CNM Barcelona to reduce implant resistance.

• The effect of the R-C biasing network at the backplane is under study and no surprises seen, yet. 19