Simulations of Radiation Damage in 3D detectors

1
Simulations of Radiation Damage in 3D detectors

• David Pennicard, Glasgow

2
Radiation damage in ISE-TCAD

• Trap models have been taken from work done at
  University of Perugia
• Most recently, Numerical simulation of radiation
  damage effects in p-type and n-type FZ silicon
  detectors, IEEE Trans. Nucl. Sci. vol. 53, pp.
  2971-2976, 2006.
• Trap models were initially based on direct
  measurements of traps from DLTS etc., then trap
  concentrations cross-sections were tuned to
  match macroscopic experimental results.

3
Reproducing results planar detectors

• The experimental results used in the paper were
  current, depletion voltage and CCE tests on
  p-type, oxygen-free FZ planar detectors
• Oxygenated detectors could give better
  performance. M. Petasecca did a presentation on
  oxygenated p-type sims at last RD50, but this
  isnt published yet.
• The physics models used consider oxide charge as
  well as bulk traps, but avalanche breakdown was
  not included.
• To start with, I worked to reproduce the results
  shown in the paper

4
Reproducing results IV / depletion
5
Reproducing results - CCE
6
Traps in full-3D detectors

• Simulated full-3D detectors up to 1016neq/cm2.
• Detectors are
• P-type substrate (71011cm-3)
• N column readout
• 55µm pitch
• 300µm thick

7
I-V results still show reasonable Vdepletion
32V
16V
6V
8
(No Transcript)
9
Charge collection simulation

• What about variation in charge collection with
  position?
• In most of the following simulations, flooded the
  entire pixel with uniform charge as a test of the
  average CCE
• Have also considered variation with horizontal
  position for one case

10
Charge collection simulation
Increasing radiation damage reduces the output
pulse, but the pulse shape isnt greatly affected
11
Charge collection simulation
Simulation predicts gt50 CCE after 1016neq/cm2
12
Horizontal position and CCE

• A series of simulations tested the variation in
  CCE with horizontal position
• Carrier drift is horizontal, and E-field varies
  with horizontal position
• Particles passing vertically through the detector
  were simulated for 15 different positions with
  both zero and 1016neq/cm2 damage (100V bias)
• With zero damage, simulations give uniform CCE
  (within 2) for each simulation
• Exception charge collection reduced within
  electrodes

13
The MIP simulations map out the pixel reasonably
well. By symmetry, the results can be extended
to the top-left section of the pixel.
14
These results show the highest CCE for MIPs
arriving around the n readout column, and a
decrease in CCE towards the p bias column.
15
Alternative 3D structures
Alternative square layout
Hexagonal layout
16
(No Transcript)
17
(No Transcript)
18
Alternative structures - CCE results
Neither alternative structure shows better
performance than the standard structure after
radiation damage. Depletion voltages are also
much the same as the standard structure
19
ATLAS 3D devices

• 3D ATLAS devices were tested for CCE after
  irradiation by Cinzia da Via et. al.
• N readout
• 230µm-thick substrate, high-resistivity n-type
• Each 50µm400µm pixel has 3 n columns, shorted
  together
• In effect, we have three elongated 50µm133µm
  sub-pixels, giving a much larger electrode
  spacing than in the simulations so far.
• Simulated this device geometry
• Used a p-type substrate in simulation not a
  match to the real detectors in the undamaged
  case, but after damage the n-type substrate will
  type invert.

20
(No Transcript)
21
ATLAS 3D devices

• Simulated depletion voltages
• Higher for this structure than in previous
  simulations, due to greater electrode spacing
• 100V for 8.61015neq/cm, compared to 35V for
  1016neq/cm for 55µm-square pixel
• CCE tests
• The bias used in the experimental tests was
  increased from 60V to 160V as the radiation
  damage increased. The biases in the CCE tests
  were chosen to match this
• For each simulation, the bias was 50V higher
  than the full depletion voltage

22
(No Transcript)
23
ATLAS 3D devices
Simulated charge collection is lower than the
experimental results In simulation, reducing the
electrode spacing improves collection greatly
24
Other simulation work 3D-STC strip detectors
25
3D-stc strip detectors

• Have done simulations of 3D-single-type-column
  strip detectors, rather than just pixels
• Simulated devices have 2 strips with strips of
  p-stop between them, matching the devices tested
  at Freiburg.

26
(No Transcript)
27
Strip 3D-stc effect of p-stop

• Change in electric field around p-stop means
  holes generated just below the surface will drift
  to the p-stop, rather than back surface

28
Simulation of laser pulses

• Freiburg tested the variation in detector
  response with position using an IR laser
• The laser pulses were approximated by a linearly
  decaying track of charge, 100µm long, moving from
  the front surface down into the substrate
• In practice, laser beam should decay
  exponentially with distance
• Since 2 strips were used, the readout currents at
  both neighbouring strips were simulated

29
3D-stc example of two signal pulses
Fast pulse from electron collection polarity
depends on laser position
Slow pulse from hole collection polarity is the
same regardless of position
30
3D-stc Charge from laser pulses
Laser pulse before mid-point of strip
Large dependence on integration time
Laser pulse arrives past mid-point of strip
31
3D-stc Charge from laser pulses- 20ns
For a short (20ns) integration time, mainly see
electron collection signal
32
Total charge seen on both strips
Missing charge matches region where p-stop
alters field pattern