Charge Collection Efficiency Studies with Irradiated Silicon Detectors

1
Charge Collection Efficiency Studies with
Irradiated Silicon Detectors

• ? Evaluation of trapping effects
• ? The influence of ballistic deficit
• ? Fits to the Charge Collection Efficiency
• ? Discussion and Conclusions

Phil Allport Gianluigi Casse Ashley
Greenall Salva Marti i Garcia
2
Evaluation of Trapping Effects

• ? Radiation damage to silicon detectors
• increases reverse currents,
• creates interface trapped charge,
• introduces traps reducing charge collection
  efficiencies
• changes the effective doping concentrations
• ? Studies of the latter effect have shown
  significant improvements under charged hadron
  irradiation when high concentrations of
  interstitial oxygen are introduced
• ? However, unlike in the case of n-side read-out
  detectors, the charge collection efficiencies for
  p-side read-out detectors do not plateau with
  voltage until well above the depletion voltage
• ? This is usually assigned to the effect of
  trapping

3
Old ATLAS Irradiated n-in-n Results
2?1014p/cm2
4
Detectors Studied

• Studies carried out with SCT miniature (cm?cm)
  p-side read-out micro-strip detectors (processed
  as test structures on ATLAS prototype wafers)
  using wide bandwidth (Phillips 6954) current
  amplifier
• Comparisons with the large area devices using
  SCT-128 analogue read-out have demonstrated good
  agreement (G. Casse et al.)

5
Evaluation of Trapping Effects

• ? Capacitance Voltage Derived Depletion Voltage

1.9?0.1?1014p/cm2 Oxygenated Miniature Micro-stri
p Detector VFD 100 ? 7 V from fitting C(V)
6
Evaluation of Trapping Effects

• ? Corresponding Charge Collection Efficiency vs
  Voltage

100 V
7
Evaluation of Trapping Effects

• ? The effects of trapping can be parameterized in
  terms of effective trapping time (Kramberger et
  al.) or, equivalently, velocity dependent
  attenuation length (Marti i Garcia et al.)
• ? In both cases, trapping is highest where the
  field is lowest
• ? These parameterizations assume timescales such
  that the total untrapped charge is collected,
  integrating over transient effects.
• ? No influence of ballistic deficit is included
• ? Both analyses give values of ? (averaged over
  e and h) that agree. ?e,h ? ?eq 1/?eff e,h
  (trapping ? flux)

8
Influence of Ballistic Deficit

• Since at the LHC electronics response times are
  close to charge collection times, signal loss due
  to incomplete charge integration must also be
  considered, which will also depend on the field
  within the detector.

Non-irradiated p-strip Detector
9
Influence of Ballistic Deficit

• ? Oxygenated and non-oxygenated detectors have
  been studied which were irradiated in the CERN-PS
  and annealed to the minimum of the Neff vs time
  annealing curve

10
Influence of Ballistic Deficit
Several miniature detectors have been studied
after irradiation and annealing.

• Detector label Fluence p cm-2 Oxygenation2
  1017 at. Cm-3
• NI Non-irradiated No
• SO1 1.9?0.1 1014 Yes
• SN1 1.9?0.1 1014 No
• SO2 2.9?0.2 1014 Yes
• SN2 2.9?0.2 1014 No
• SO3 5.1?0.4 1014 Yes
• SN3 5.1?0.4 1014 No

11
Influence of Ballistic Deficit
Miniature detectors irradiated to 1.9?1014p/cm2
Non-Oxygenated
Oxygenated
12
Influence of Ballistic Deficit
Miniature detectors irradiated to 2.9?1014p/cm2
Non-Oxygenated
Oxygenated
13
Influence of Ballistic Deficit
Miniature detectors irradiated to 5.1?1014p/cm2
Non-Oxygenated
Oxygenated
14
Non-Oxygenated Detectors Relative Ballistic
Deficit
Non-irradiated
1.9?1014 p/cm2
5.1?1014 p/cm2
2.9?1014 p/cm2
Oxygenated
15
Oxygenated Detectors Relative Ballistic Deficit
2.9?1014 p/cm2
1.9?1014 p/cm2
5.1?1014 p/cm2
16
Fits to the Charge Collection Efficiency
The above results suggest that, particularly at
high doses, the ballistic deficit is not a major
factor for LHC speed operation In the following
fits, sufficient integration time has anyway been
allowed such that the only charge loss is due to
trapping Free parameters attenuation length
?, depletion voltage VFD total generated
charge Q0
1.9?1014 p/cm2
17
Fits to the Charge Collection Efficiency
2.9?1014 p/cm2
5.1?1014 p/cm2
18
Fits to the Charge Collection Efficiency
     
19
Fits to the Charge Collection Efficiency
     
The fitted values of VFD agree with each other
and with oxygenated data from RD48 (CERN LHCC
2000-009) taking account of the proton damage
factor The fitted values of Q0 18.1?0.3,
18.2?0.3, 17.7?0.3, 18.1?0.6, 18.2?0.4 and
18.3?0.4 are all consistent and agree with the
pre-irradiation value 17.9?0.3
20
Fits to the Charge Collection Efficiency
     
The dependence of Q/Q0 and therefore ? on ? leads
to a value of ?eff 5.6?0.6?10-16cm2/ns Assuming
this value allows extrapolation of CCE to high ?
21
Discussion and Conclusions
Because for p-strip read-out, the trapping
significantly affects the CCE(V), the
improvements in VFD due to oxygenation do not
give correspondingly large effects in terms of
CCE The trapping dependence on the field leads
to CCE(VFD) being higher for non-oxygenated than
oxygenated detectors by 5 Read-out from the
high-field n-side gives less dependence on
trapping leading to the CCE(V) ? ?V behaviour
below VFD This would imply that for high doses,
n-side readout should benefit more from
oxygenation of the substrate
22
Discussion and Conclusions
The LHC-b vertex detector is proposed to use
oxygenated n-strip detectors for which the first
prototypes have just been delivered to Liverpool
and are currently being irradiated in the CERN PS
23
Discussion and Conclusions
LHC-b uses back-to-back thinned disks for r and ?
plus double-metal routing
24
Discussion and Conclusions
LHC-b and the pixel systems of ATLAS and CMS need
to maximise their survival n-side readout
oxygenated detectors look to offer the best
possibilities Super-LHC with factor of 10
increased luminosity could also need such
technology
25
Discussion and Conclusions
Detectors produced with n-side read-out do suffer
from the disadvantage of requiring potentially
expensive double-sided processing Use of p-type
substrates does provide a viable alternative
where cost is of paramount importance
Comparison of p-type and n-type detectors after
3?1014 p/cm2