Charge Collection in Single-Column 3D Detectors

Carlo Tosi, Mara Bruzzi

INFN and University of Florence

Monica Scaringella, Hartmut F.-W. Sadrozinski, Alek Polyakov

SCIPP, UC Santa Cruz

M. Boscardin, C. Piemonte, A. Pozza, S. Ronchin and N. Zorzi

ITC-irst

G.-F. Dalla Betta

DIT, Università di Trento

Analysis of CCE - V on 3D sensors

SensorsP-type strip-like Sensors ~500m thick, ~150 m single columns (n+/p)

Depletion Pattern on 3D:C-V and Cint-VConsequences for CCE

2 CCE systems: Beta source 90Sr with scintillator trigger Analog DAQ (Firenze)

2 sec shaping time, pad sensor Binary DAQ (Santa Cruz)

100 ns shaping time: strip sensor

Cint-V 1MHz

5

6

7

8

9

10

11

0 20 40 60 80 100 120 140 160 180Bias [V]

Cin

t [p

F]

Depletion Pattern on 3D SSD: C-V and Cint-V

1/C2 (10 kHz)

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

0 20 40 60 80 100 120 140 160 180

Bias [V]

1/C

2 [1

/pF2 ]

Voltage Range 2:region between col.is fully depleted depletion proceedsonly towards the back(almost like a planar diode)

full depletion ~200Vdepletion width of ~150+350m

Voltage Range 1:region between col.is not fully depleted large capacitance

full dep. between columns~ 7V

Depletion Pattern CCE PatternMean collected charge & Inefficiency

Voltage Range 2 (bias > ~7 V):region between columns is fully depletedDepletion depth = tw

-> Charge collected = 4fC*(tw/300m)-> No Inefficiency

tc

tw

a b

Voltage Range 1 (bias < ~7V): there are lateral depleted regions near each column which are tc in depth. Charges generated by traces like b are not collected.-> Charge collected = 4fC*(tc/300m) ~ 2fC-> Inefficiency = fraction of un-depleted area

between columns

Analog DAQ (Firenze) : Pad Sensor

Inefficiency(ideally, not measured here!)

Most probable Charge

Inefficiency = fraction of undepleted areaMean charge ~ depleted thickness

Pad Sensor Analysis Analog DAQ (Firenze) with deconvolution software

(NIKHEF)

150V

Discrepancy between pulse height and simulated Landau is observed at high voltage. Most probable value of pulse height is determined by Landau deconvolution only in the range 0-60V. At 70-150V the most probable value taken is max of pulse height spectrum.

Applied Voltage = 5V

Pulse Height Max

mp from Landau deconv

0 50 100 15020

40

60

80

100

120

max

pul

se h

eigh

t [a.

u.]

Voltage [V]

Fz Si, W = 525m

Sc-3D D4 diode configuration

Efficiency vs. Threshold

0

0.2

0.4

0.6

0.8

1

0.5 1 1.5 2 2.5 3 3.5 4 4.5Charge (fC)

Eff

icie

ncy

6 V

0 V

30 V

Binary DAQ (Santa Cruz) Strip Sensor

(We define the median as the collected charge)

At 6V:Collected charge = 1.7 fCAs expected from 140 m Inefficiency = 0i.e. un-depleted area = 0

At 0V:Collected charge ~ 1fCInefficiency > 80%?

At 30V:Collected charge = 3.0 fCAs expected from 280 m Inefficiency = 0i.e. un-depleted area = 0

?

Median charge

Most likely charge

Reconstruction of Pulse Height Spectrum

Mean Charge and Inefficiency

Efficiency vs. Bias Voltage

0

0.2

0.4

0.6

0.8

1

0 5 10 15 20 25 30

Bias Voltage (V)

Eff

icie

nc

y

Collected charge:Increases with bias voltage(Square root behavior beyond ~7V?Not yet saturated at 150 V)

Efficiency varies only below ~ 7V,~ 100 % beyond.

CCE in p-type 3D SSD

0

1

2

3

4

5

6

0 50 100 150 200

Bias Voltage [V]

Co

llec

ted

Ch

arg

e [

fC]

Median

Most Likely

Binary DAQ (Santa Cruz)

Collected charge: compares well with Ansatz Q = 1fC + const/C

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

0 20 40 60 80 100 120 140 160

Voltage (V)

Ch

arg

e (

fC)

Median Charge (fC)

Expedted Q from 1/C

Charge Collection Comparison: Pads vs. SSD, 100 ns vs. 2.4 s

All data normalized at 150 V bias.

Good agreementbetween UCSC (100ns) and FI (2.4s).

Higher active thickness than from C-V i.e. d ~ 1/C (diffusion?)

0

100

200

300

400

500

600

0 50 100 150 200Voltage [V]

Act

ive

Th

ickn

ess

[um

]

Florence

C-V

UCSC

ConclusioConclusionsns

Charge Collection vs. voltage confirms simple picture of depletion in single-column 3D sensors:

• Rapid depletion between columns (< 10 V)• Slow, planar-diode like depletion beyond that• Voltage dependence of charge collection about same for SSD (100 ns) and pads (2.4 s)• Collected charge larger than predicted from 1/C at lower voltages

Ackgnowledgments::

ITC-irst team for production of 3D detectorsClaudio Piemonte for the very nice graphics of depletion process