Measurement of the Trapping Time Constants in Silicon with...
Transcript of Measurement of the Trapping Time Constants in Silicon with...
Measurement of theTrapping Time Constants in Silicon
with the Transient Current Technique
*O. Krasel, C. Gößling, J. Klaiber-Lodewigs, R. Klingenberg,M. Maß, S. Rajek and R. Wunstorf for the ATLAS Collaboration
Lehrstuhl f. Experimentelle Physik IV, Universität Dortmund
Olaf Krasel, Universität Dortmund 2
• fluence up to 1015 1-MeV-neq/cm2
for design luminosityData on charge collection efficiency are needed for:• detector simulation• optimization of operation conditions
Motivation: The ATLAS Pixel Detector• innermost part of tracking system
• 3 points for tracking
• coverage 0≤η≤2.5
• 3 barrel layers, 6 wheels
• total active area ≈2m² (1744 modules with ≈80 million pixel)
Olaf Krasel, Universität Dortmund 3
Effective Trapping Time
• Trapping leads to charge carrier loss:
• charge induced on electrodes by driftingcharge (Ramo’s theorem):
• resulting (measured) signal current:
• injection with short range laser from one sideallows to distinguish between electron andhole signal
)(with ,)(1
)( eqeffeffeff
dttQtdQ Φ=−= τττ
)/exp()()( 0effm ttv
dQ
ti τ−=
dttvdQ
dxdQ
dq )(==
Olaf Krasel, Universität Dortmund 4
Transient Current Technique, Set-up
• 672nm red laser (3.5µm absorption length, FWHM = 44ps),• voltage source with picoamperemeter (Keithley 487)• fast pulse amplifier (10×, 100 kHz - 1.8 GHz), (current sensitive!)• oscilloscope (Tektronix TDS 784D, band width 1 GHz)• rise time of system (incl. detector) about 1 ns• PC readout system (LabVIEW)• cooling system (-20°C - +20°C, RMS 0.2°C)
Olaf Krasel, Universität Dortmund 5
Charge Correction Method (CCM)• correction of measured current:
•
• t obtained from slope vs. t plotdepbias
cc
VV
dttiQ
>
== ∫ and correct for
const.)(
τ
( )τ/exp)()( ttiti measc +⋅=icorr(corrected)
im (measured)
lines with τ=const.
Olaf Krasel, Universität Dortmund 6
Exponentiated Charge Crossing (ECC)
dtttiQ meas )/exp()(exp τ+= ∫• exponentiated charges from different Vbias are plotted vs. 1/τ• 1/τ is obtained from mean of intersection points of lines
icorr(corrected)
im (measured)
lines with Vbias=const.
Olaf Krasel, Universität Dortmund 7
Samples
• irradiated with 24 GeV protons at CERN-PS,no biasing during irradiation
• 4 diodes from ATLAS-Pixel pre-production (CiS),irradiated to 1.1, 3.2, 5.0 and 8.9⋅1014 neq/cm2
annealed to minimum in Vdep at 60ºC• 2 diodes from SRD-Project (CiS),
irradiated to 4.5 and 6.5⋅1014 neq/cm2;annealed to minimum in Vdep at room temperature
• ⟨111⟩ crystal orientation,24h at 1200ºC oxygenation
For the irradiation of the samples I want to thank M. Glaser,M. Moll (CERN) and P. Sicho (Acadamy of Sciences, Prague)
Olaf Krasel, Universität Dortmund 8
Trapping Times: Fluence Dependence
/nscm10)16.016.5( 216−⋅±=eβ /nscm10)16.004.5( 216−⋅±=hβ
• errors include only “statistical” error, systematics not yet included• comparison with results obtained by other groups (extrapolated from data taken at fluences < 2.4⋅1014neq/cm2): agreement for electrons, deviation for holes
Olaf Krasel, Universität Dortmund 9
Trapping Times: Annealing
holes,cm/1046.4 2eq
14eq n⋅=Φ
electrons
,cm/1046.4 2eq
14eq n⋅=Φ
( ) ( )( )( )( ) 19% by increase h,59.069.1
20%,by decreaseh,80.077.2
/exp1/exp)(
,
,
0
≈±=
≈±=−−⋅+−⋅= ∞
ha
ea
aa ttt
τ
ττβτββ
• ansatz for annealing function:
Olaf Krasel, Universität Dortmund 10
Charge Collection Efficiency
electrons
,cm/1046.4 2eq
14eq n⋅=Φ
holes
,cm/1046.4 2eq
14eq n⋅=Φ
• CCE changes during annealing due to changes in Neff and t
• charge collection efficiencies CCEmin refer to chargesinjected close to surface
• therefore they give only a lower limit for CCE interesting inHEP experiments
Olaf Krasel, Universität Dortmund 11
Charge Collection Efficiency
• CCEmin for electrons is40-50% at Vdep50-60% at Vdep+50V (as foreseen in ATLAS)
electrons
,cm/1046.4 2eq
14eq n⋅=Φ
Vdep from chargecollection curve,? Vdep=15V
Olaf Krasel, Universität Dortmund 12
Conclusions• the effective trapping times of holes and electrons in
silicon have been measured for fluences between 1.1 and8.9·1014 1-MeV-neq/cm2
•
• confirmed• results for electrons are consistent with extrapolation of
measurements at lower fluences (Brodbeck et al., 2000;Kramberger et al., 2000)
• trapping times for holes are not consistent with thesemeasurements
• annealing of t is negligible, for electrons it even improvesslightly (important for n+-on-n-pixels)
( ) ( )
)ts!measuremenother to(contrary
ns,/cm1016.004.5 ns,/cm1016.016.5
e
216216
h
he
ββ
ββ
≈
⋅±=⋅±= −−
eqΦ⋅= βτ1
Olaf Krasel, Universität Dortmund 13
Finally...
Olaf Krasel, Universität Dortmund 14
Signal Treatment
• bandwidth:
• tail (problems withreflections/oscillations):
DRC
BW
mRCmm
RC
tu
tututu
==
+→
τ
τ)(
)()()( &
parametersfit are,
point inflection
for for
)/exp()(
)(
0
0
f
ip
ip
ip
f
BWBW
tu
t
tttt
ttutu
tu
=
>≤
−→